# Research Papers — 2026-09 · bundle page 128 of 266

> Full-text markdown of papers #6351–6400 of 13298 published in 2026-09, newest first. AI-readable mirrors of the full 13298-paper 2026-09 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-09/page/128
[← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/127.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-09/page/129.md)

### Robust Market Making with Hawkes Order Flow and Price Impact via Adversarial Reinforcement Learning

Authors: Hao Yang, Zhenguo Xu
Published: 2026-09-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2609.22785v1

Market-making strategies in real limit order book markets face substantial model uncertainty and regime-shift risk. Existing adversarial reinforcement learning approaches improve robustness by formulating the Avellaneda--Stoikov market-making problem as a zero-sum game between a market maker and an environmental adversary. However, these approaches typically rely on Poisson order arrivals and neglect trade-induced price impact, limiting their ability to capture important high-frequency market microstructure effects such as clustered order flow, self-excitation, and post-trade price feedback. We extend adversarial reinforcement learning for market making to a more complex environment with Hawkes self-exciting order arrivals and trade-induced price impact. To mitigate the increased non-stationarity introduced by the expanded regime space, we incorporate an LSTM module that explicitly models the temporal structure of recent observations. We further characterize the equilibrium properties of the proposed framework through both game-theoretic analysis and numerical experiments, and introduce a robustness evaluation protocol focused on improvements in the left tail of the return distribution. Experimental results across a range of market regimes show that the proposed method achieves improved left-tail performance in most complex microstructure environments. In particular, the gains are pronounced in regimes with strong Hawkes excitation and low-to-moderate price impact. Bootstrap tests provide no evidence that these improvements are obtained through a stronger terminal directional inventory bias. These results suggest that combining adversarial training with temporal state representation can improve the robustness of reinforcement-learning-based market-making strategies under order-flow self-excitation, price impact, and regime uncertainty.

---

### Improved Private Sparse Covariance Estimation with Multiscale Threshold Tests

Authors: Zihan Zhang
Published: 2026-09-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2609.22783v1

We study differentially private covariance estimation in operator norm for mean-zero sub-Gaussian distributions with unknown covariance support and at most $k$ nonzero entries per row. We develop a multiscale random-threshold algorithm with sample complexity $\ot(k^2/α^2+k\sqrt d/(α\varepsilon))$ for $(\varepsilon,δ)$-differential privacy and error at most $ασ^2$, where $d$ is the dimension and $σ$ is a known sub-Gaussian scale. The bound improves the privacy-dependent term of the existing $\ot(k^2/α^2+k^{3/2}\sqrt d/(α\varepsilon))$ \citep{kumar2026curse} upper bound by a factor of $\sqrt k$, and matches the lower bound of $\widetildeΩ(k^2/α^2 + k\sqrt{d}/(α\varepsilon))$ in its applicable parameter regime. Our key technical ingredient is a direct operator-norm bound on the centered fluctuations of an ideal reconstruction, exploiting conditional independence rather than accumulating entrywise errors across each row. A multiscale allocation of threshold tests balances reconstruction variance against query sensitivity. Together, these ingredients sharpen the trade-off between approximation error and privacy protection, removing the additional $\sqrt{k}$ factor from the privacy-dependent sample complexity.

---

### Look Before You Steer: Geometry Predicts SAE Feature Steerability

Authors: Muhammad Khan, Shlok Channawar, Akshaj Gurugubelli, Girish Gupta, Aditya Shah
Published: 2026-09-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2609.22782v1

Steering with SAE features requires per-feature coefficient tuning, which currently demands intervention sweeps. We ask whether properties of the SAE itself, computable before any forward pass, predict which features will be cheap or expensive to steer. We show that variation in SAE feature steerability is partially predicted by decoder-space geometry: neighbor density and maximum cosine similarity to nearby decoder directions, both computable from the SAE weight matrix before any intervention, rank features by how much steering they require for a fixed behavioral effect ($ρ$ up to $-0.546$, $p < 10^{-6}$, AUROC 0.610-0.822 across conditions; the signal is rank-based, consistent with grid discreteness). This geometry-steerability relationship replicates across two Gemma-2 model scales (2B and 9B), two SAE widths (16K and 65K), and is detectable cross-architecturally on Llama-3.1-8B-Instruct ($ρ= -0.266$, $n = 300$). On Qwen3-8B with BatchTopK SAEs, geometry predicts whether a feature is steerable at all but not the continuous ordering among responsive features, revealing a boundary condition tied to SAE training regime. The signal weakens at deep proportional layer depth in both models, where the cost of steering exceeds our intervention budget, a consistent depth boundary. These results provide preliminary evidence that pre-steering geometry can partially inform coefficient selection, offering a path toward screening features for controllability before deployment.

---

### MIS-Bench: Benchmarking Multimodal LLMs for Psychotherapeutic Interpersonal Skills Assessment

Authors: Yuhan Lu, Yi Yao, Hua Shen, Katie Aafjes-van Doorn, Zhaonan Wang
Published: 2026-09-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2609.22778v1

Multimodal large language models (MLLMs) are increasingly used as evaluators, yet their reliability in professional assessment tasks that require expert judgment remains unclear. We investigate this challenge in the context of assessing psychotherapeutic interpersonal skills and introduce MIS-Bench, a Multimodal Interpersonal Skills (MIS) benchmark comprising 996 psychotherapy response videos annotated across 8 dimensions of Facilitative Interpersonal Skills. Across 9 MLLMs with multiple modality and prompting settings, we find that current models show only modest agreement with human experts, inconsistent gains from multimodal input, and limited benefits from reasoning-based prompting. To mitigate this gap, we propose MIS-RAFT, a regression-aware fine-tuning method inspired by RAFT and tailored to fine-grained interpersonal skill scoring at one-decimal precision. MIS-RAFT addresses the mismatch between autoregressive token prediction and scalar-valued expert assessment, significantly improving agreement with human ratings. Overall, MIS-Bench reveals a clear gap between general multimodal capability and expert-level interpersonal judgment, while MIS-RAFT offers a promising path toward more reliable model-based assessment.

---

### DVA-Neurons: Design and Verification of Adaptive LIF Neurons: From Single-Neuron Dynamics to Multi-Neuron Spiking Networks

Authors: Thanh Pham, Riadul Islam
Published: 2026-09-19
Categories: cs.AI, cs.AR, eess.SP
arXiv: https://arxiv.org/abs/2609.22775v1

Spiking Neural Networks (SNNs) offer a promising path toward ultra-low-power artificial intelligence inference by emulating the event-driven computation of biological neurons. However, two challenges limit their practical deployment. First, fixed-parameter Leaky Integrate-and-Fire (LIF) neurons lack the adaptation mechanisms observed in biology, where neurons modulate their excitability based on firing history. Second, scaling from single neurons to multi-neuron networks introduces challenges in synaptic weight distribution and inter-neuron spike routing that are absent in isolated designs. This paper addresses both issues through the extension, verification, and physical implementation of adaptive LIF neurons at three architectural scales. This work contributes: a 2nd-order neuron with two-stage synaptic filtering for richer temporal dynamics; a fully-connected 6-neuron spiking network with configurable weights (100 to 5) demonstrating weight-based inter-neuron communication; and a direct verification methodology enabling per-cycle observation of all internal states. All designs were synthesized targeting Selected Area Electron Diffraction (SAED) 14 nm Complementary Metal-Oxide-Semiconductor (CMOS) technology at 1 GHz and verified with Cocotb-based Python testbenches under pulsed current stimuli (amplitude 80, ISI=3). The results show that adaptation effectively modulates firing: 31% suppression in the 2nd-order neuron (25 vs.\ 36 spikes) and 31% reduction in postsynaptic firing in the network (18 vs.\ 26 spikes). Physically, the 2nd-order neuron costs 1.77x more area and 1.52x more power than the 1st-order baseline, while the 6-neuron network demonstrates near-linear scaling (5.7x area, 5.3x power). Seven verification bugs spanning testbench connectivity, fixed-point overflow, and Verilog expression-width semantics are documented.

---

### NLPCC 2026 Task 10: Citation-Level Faithfulness Verification with DeBERTa Ensembles and Class-Wise Calibration

Authors: Yanling Li, Zirui Li, Mingyu Wan
Published: 2026-09-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2609.22774v1

This paper presents our system for Track 2 of the NLPCC 2026 Shared Task 10 on citation-level faithfulness in AI-assisted scientific reporting. Given an atomic scientific claim and the structured full text of its cited paper, the task requires both a four-way relation label and up to three evidence paragraph identifiers. The label head ensembles a paragraph-aware cross-encoder with a document-level DeBERTa-large classifier, followed by class-wise decision calibration. Probability-level fusion is motivated by an out-of-fold tendency to over-predict Topical Match. The evidence head combines paragraph scores from top-20 and top-30 joint models with BM25 scores. The system runs fully offline without external retrieval or LLM prompting. On the final leaderboard, our system achieved 82.9898 overall (89.5491 Macro-F1 and 76.4305 Joint@3), ranking second in Track 2. Ablations and error analysis show that model complementarity and calibration drive the label gains. Gold-evidence inference changes label Macro-F1 negligibly, whereas evidence ranking remains important for Joint@3.

---

### ParA-LLM: A Unified Approach to Paralinguistic and Acoustic Speech Understanding

Authors: Nishit Anand, Jiaqi Su, Ke Chen, Yunyun Wang, Dinesh Manocha, Ramani Duraiswami, Rithesh Kumar, Zeyu Jin
Published: 2026-09-19
Categories: cs.SD, cs.AI, cs.LG, eess.AS, eess.SP
arXiv: https://arxiv.org/abs/2609.22771v1

Recent advances in Audio LLMs have achieved human-level speech recognition, yet existing systems struggle to capture paralinguistic aspects such as speaker traits, expressive variations, and environmental acoustic conditions. To address this, we design a framework of 22 paralinguistic characteristics and create a dataset of over 1.2M Audio-QA pairs. We develop ParA-LLM, trained with a two-stage curriculum: first on single-attribute questions to build foundational knowledge, then on multi-attribute questions for joint reasoning over speaker and acoustic characteristics. We also release ParA-Bench, a benchmark of 6,000 multiple-choice questions across speaker-speech, acoustic, and mixed categories, where frontier models like GPT-4o-Audio achieve only 36% accuracy. ParA-LLM surpasses state-of-the-art Audio LLMs like GPT-4o-Audio by 7.5% on ParA-Bench, with additional gains of 1.13% on MMAU-Pro Speech and 7.49% on MMAR Speech.

---

### Parameterized Dense-Sparse Fusion for Hybrid Retrieval: Tuning a Rank-Score Mix on BEIR SciFact with Qdrant

Authors: Satyanarayan Pati, Srikanth Patil
Published: 2026-09-19
Categories: cs.IR, cs.AI
arXiv: https://arxiv.org/abs/2609.22770v2

We study a parameterized hybrid ranker that fuses a dense embedding list and a sparse lexical list. The method has a small, explicit parameter vector: a dense prior $α\in (0,1)$, a score-versus-rank mix $λ\in (0,1)$, an RRF smoothing parameter $κ> 0$, optional list-geometry coefficients that move $α$ per query, and a router margin $τ$ that can turn sparse search off. We grid-search those ranges on SciFact train (809 queries) and freeze the chosen values on SciFact test (300). The tuned rank-score mix ($α= 0.8$, $λ= 0.75$, $κ= 20$) reaches 0.753 nDCG@10 and 0.889 recall@10, outperforming dense BGE (0.742 / 0.871) and equal-weight RRF (0.707 nDCG@10) on that test split. A list-conditioned $α$ adds +0.0006 nDCG; a sparse-off router is rejected by the same train split (any $τ$ that skipped approximately 50% of queries lost nDCG). These coefficients are dataset-specific. Equal RRF with the same models does not beat dense on a nine-zip BEIR macro-average (0.479 vs. 0.519 nDCG@10). Repeating the same train-then-freeze sweep independently on all 20 indexed units beats equal RRF on 20/20 and dense on 16/20 (unit-mean nDCG@10 0.467 vs. 0.462 dense vs. 0.420 RRF). Other corpora should reuse the ranges, not a copy of the SciFact point.

---

### Beyond Final-Token Classification: Heterogeneous Readouts for Evidence-Grounded Suicide Risk Detection

Authors: Zirui Li, Yanling Li, Kaolanglang Gao
Published: 2026-09-19
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2609.22767v1

The IEEE BigData Cup benchmark combines three prediction problems with different output structures: ordinal suicide-risk classification, multi-label psychosocial factor detection, and extraction of supporting phrases. We introduce heterogeneous readout decomposition (HRD), which separates semantic verification from output realization. A locally deployed Qwen3.8-27B model, adapted with task-specific QLoRA adapters, produces both answer-token margins and layer-63 answer states for card-conditioned queries. HRD compares four latent scores for ordinal risk, retains token margins for most factors while routing seven labels through one shared latent probe, and constructs evidence sets from verbatim span candidates with calibrated, risk-conditional constraints. On two held-out user-grouped confirmation folds, the latent risk readout improves weighted F1 from 0.8237 to 0.8372 and macro F1 from 0.7965 to 0.8185. Selective factor routing improves macro F1 by 0.0105 and tail-label macro F1 by 0.0189; in contrast, global latent replacement and independent label-specific probes fail. The constrained evidence decoder raises pooled phrase F1 from 0.7488 to 0.7609 in row-level out-of-fold evaluation. Lenormand's best public result is 0.8052 on Subtask 1 and 0.6636 on Subtask 2, giving a composite score of 0.7627. These results identify the answer-token boundary, rather than semantic representation alone, as a measurable source of error in this benchmark.

---

### DriveReferee: Geometric Safety Verdicts Need Not Be Learned for Driving World-Action Models

Authors: Fengcheng Yu, Dhruv Parikh, Junjie Ye, Maulik Bhatt, Thang Vu, Igor Vasiljevic, Vitor Guizilini, Yue Wang
Published: 2026-09-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.22762v1

Generative world-action models (WAMs) jointly generate future video and vehicle actions, while their action branches remain primarily optimized by expert imitation. Yet imitation provides no explicit closed-loop geometric verdict for generated trajectories, making verification important during both training and deployment. Closed-loop evaluators can check collision and drivable-area violations, but require privileged scene state unavailable at deployment. Existing approaches often close this gap by learning a verifier from sensor features. For these geometric checks, the rule itself is explicit. For example, collision is determined by whether the rolled-out ego footprint overlaps occupied vehicle space. What is unavailable at deployment is the scene state needed to apply the rule. We introduce DriveReferee, which uses a learned geometry readout to predict the scene representation from camera observations and executes the geometric safety rule directly rather than learning it. The resulting analytic referee evaluates collision and drivable-area safety from a scene state and candidate trajectory. During training, it scores self-sampled trajectories on ground-truth state and distills the resulting preferences into the WAM policy. At deployment, the same referee evaluates generated trajectories on this predicted state and selects a safer alternative when needed. The analytic referee requires no verdict-specific training, and its decisions follow an explicit geometric rule. Under matched candidates and inference budgets, it matches or outperforms all learned-verifier and heuristic baselines. Given the same predicted state and trajectory, learning the verdict provides no measurable downstream gain despite requiring tens of thousands of evaluator-labeled training examples. On the full NAVSIM navtest, DriveReferee reaches 92.02 PDMS with single-camera visual input and no external training data.

---

### CTSpinoPelvic1K: spine, pelvis, ribs and femora in one coordinate frame, annotated for lumbosacral transitional anatomy

Authors: Gregory Schwing, Ashley Schehr, Annika Tekumulla, Margret Khoushi, Ryan Christian, Dane Hubers, Faris Mahjoub, Hassan Saad, Mia Sooch, Sathyagopal Siddapureddy, Michael McLellan, Jerick Kim et al.
Published: 2026-09-19
Categories: cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2609.22760v1

Purpose: A vertebra at the lumbosacral junction is named by counting caudally from C2 on whole-spine imaging, but a lumbar case is planned on lumbar-only imaging (T12 to S1), without C2. Abdominopelvic CT holds that span plus the lowest ribs and pelvis. Where a lumbosacral transitional vertebra (LSTV) alters the count, the local anatomy is ambiguous: four rib-free vertebrae may be an L1 with a lumbar rib or an L5 assimilated to the sacrum, and six may be a sixth lumbar vertebra, a T12 with aplastic ribs, or a lumbarized S1. CTSpinoPelvic1K asks whether local morphology resolves it without the count. CTSpine1K's vertebrae and CTPelvic1K's pelvis covered these patients but were never joined; this release joins them on one series and adds the bones neither had. It provides 802 CT records with per-level ribs and femora, levels anchored on the lowest rib-bearing vertebra and S1, and classes for L6, T13, a separate S1 and lumbar ribs, so anomalies are recorded as such. Acquisition and Validation Methods: Records pair CTSpine1K and CTPelvic1K labels on each patient's bone-richest series under a VerSe-native scheme. Validation covered geometric invariants (802/802 pass), rib-vertebra incidence across 5,749 ribs (0.035% offset), and spinopelvic measures matching published values. Data Format and Usage Notes: NIfTI image/label pairs with patient-grouped LSTV-stratified five-fold splits and a loader; archived at https://doi.org/10.5281/zenodo.22139642. Potential Applications: Classifying a vertebra from local features; updating cadaveric morphometry; spinopelvic assessment; opportunistic screening; and, absent a public preoperative lumbar cohort, surgical planning research (377 records prone). Limitations: thoracic ground truth is field-of-view limited; postural angles supine; no held-out test set; ribs are triaged-review pseudolabels; Castellvi grades two-reader consensus on 33 records.

---

### Algorithmic Collusion and the Complexity of Information-Value-Free Equilibria

Authors: Ioannis Anagnostides, Weiqiang Zheng
Published: 2026-09-19
Categories: cs.GT, cs.CC, cs.DS, cs.LG
arXiv: https://arxiv.org/abs/2609.22757v1

A (coarse) correlated equilibrium (CE) is information-value-free (IVF) if a player can match the payoff obtained from recommendations by committing to a fixed action. Motivated by the problem of regulating algorithmic collusion, this refinement was introduced by Hartline, Wang, and Zhang (EC'26), who showed that it can be computed in polynomial time in explicitly represented normal-form game. In this paper, we examine the complexity of IVF(C)CEs in succinct games, which model more realistic strategic interactions that feature either many players or exponentially many pure strategies. We first show that computing an information-value-free CE is PPAD-complete in many-player polymatrix games or two-player Bayesian games, even when the approximation is a constant. We also prove an unconditional exponential query lower bound. Our results establish that IVFCEs are intractable, even in the centralized model, and rule out the existence of any efficient learning dynamics. This significantly strengthens the impossibility result of Hartline, Wang, and Zhang, which concerns a particular class of learning algorithms, and furnishes strong computational critiques of recent regulation on algorithmic collusion. To sidestep these hardness results, we examine the complexity of information-value-free CCE. Certain no-regret algorithms---such as regret matching or FTRL---provide a fully polynomial-time approximation scheme (FPTAS) for this problem. The complexity when the approximation is exponentially small turns out to be nuanced. On the one hand, leveraging no-regret dynamics, we establish membership in $\text{CLS} = \text{PPAD} \cap \text{PLS}$. On the other hand, we show that it is at least as hard as the P-matrix linear complementarity problem, and hence as hard as simple stochastic games. This shows that even IVFCCEs are unlikely to admit a polynomial-time algorithm barring a major breakthrough.

---

### D-IMPL: A Diffusion-based Solver for Parameterized BBOs

Authors: Yang Hu, Na Li
Published: 2026-09-19
Categories: cs.LG, math.OC
arXiv: https://arxiv.org/abs/2609.22752v1

Diffusion models have demonstrated strong power in generative modeling tasks across multiple domains, exhibiting a remarkable capability of learning complex distributions from samples. In this paper, we leverage such capability to design an efficient universal diffusion-based solver for parameterized black-box optimizations (BBO), where the optimizer has only black-box access to queries of the objective function at the learning stage, yet is able to reduce the additional computational cost at the inference stage for each BBO instance while also capturing the potential multi-modal landscape of non-convex objectives. To cast our formulation as a compatible generative modeling task, we introduce the notion of minimization policy as a new solution concept, which defines a sampling distribution over the solutions that should concentrate around the minimizer set for each BBO instance. We then propose Diffusion-based Iterative Minimization Policy Learning (D-IMPL), a practical generative-model-based solver for solving parameterized BBOs that employs diffusion models to learn a minimization policy, whose density is proportional to the exponential of the negated objective values, thereby amortizing the computational costs across different BBO instances. Furthermore, we demonstrate the performance of our D-IMPL algorithm by establishing a sample complexity guarantee showing that a $δ$-approximate minimization policy can be effectively learned within $O(\log(1/δ))$ iterations, and by extensive empirical evaluations over a range of constrained and unconstrained BBO tasks.

---

### Towards Robust Classroom Attendance: A Comprehensive Evaluation of Face Detection and Recognition Models

Authors: Himani Trivedi, Hiren Patel, Ridham Patel, Krutika Patel, Nancy Patel
Published: 2026-09-19
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2609.22750v1

Manual attendance methods, such as paper or register-based systems, take a lot of time, can lead to errors, and are easy to falsify. Face recognition is more reliable, but it frequently struggles in classrooms because lighting and other conditions can vary. Face recognition datasets are designed for regulated environments and do not capture the actual challenges found in classrooms. To address this, a new face detection and recognition dataset, the Visage Face dataset, comprising 16,234 face samples, is proposed for the task of face detection and recognition. The photos are taken from different angles and under varying lighting conditions, with students showing a range of expressions, and some faces partly covered to reflect real-life situations. A YOLO-based system is used to detect faces and tested seven advanced face recognition models with thirteen configurations: LVFace, QCFace, FaceLiVTv2, TopoFR, EdgeFace, TransFace, and GhostFaceNets. Of these, FaceLiVTv2-M performed best, with 99.75% Top-1/Top-5 accuracy and an inference time of 6.459 ms. These results show that the Visage Face Dataset is a realistic and challenging benchmark for face recognition in classroom attendance.

---

### Beyond Raw Engagement: A Counterfactual Observability Framework for Recommender Systems at Netflix

Authors: Chaoran Guo, Ding Tong, Ting-Po Lee, Scarlet Chen
Published: 2026-09-19
Categories: cs.IR, cs.AI
arXiv: https://arxiv.org/abs/2609.22747v1

Understanding the performance of large-scale recommender systems remains an underexplored challenge, especially for content creators and model developers. The raw engagement signals available to them, such as views and clicks, conflate content quality, model behavior, presentation bias, and audience reach, making it hard to attribute outcomes to the right cause. In this work, we present a general evaluation framework that enhances observability across multiple recommender systems at Netflix and demonstrate its effectiveness through several production deployments. The framework treats recommender-system observability as a counterfactual measurement problem: estimating what the recommender would have done, and what engagement would have followed, in the absence of a specific content item or model decision. We articulate three stakeholder-centered observability principles for content creators and model developers, and propose measurement methodologies covering bias reduction, relativity, and incrementality, applicable to both single-stage and cascading recommender systems and serving both audiences from a single measurement foundation.

---

### ProcessLight: Process Supervision for Large Language Model Based Traffic Signal Control

Authors: Huaitao Zhao, Tianlong Zhou, Weijie Wang, Jiasheng Shi, Weixiong Rao
Published: 2026-09-19
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.22746v1

Large Language Models (LLMs) have recently been introduced into traffic signal control (TSC) as decision agents due to their strengths in human-readable reasoning generation. Yet, existing LLM TSC methods optimize only from final outcomes and fail to distinguish valid from flawed reasoning steps, causing useful or misleading steps to be jointly updated and thus impairing the model's learning of effective reasoning. To bridge this gap, we propose an LLM-based framework ProcessLight to decompose signal decisions into verifiable semantic steps. Building on ProcessLight, we further develop Step-wise Traffic Process Policy Optimization (STeP-PO), a novel reinforcement learning framework that optimizes structured reasoning processes through step-level credit assignment. Specifically, STeP-PO uses step quality scores to evaluate local reasoning quality and step importance to measure each step's influence on the final action, and then assigns step-level advantages over a semantic step tree structure. The resulting step-level advantages are propagated to reasoning tokens, enabling fine-grained policy optimization beyond outcome-only rewards. Extensive experiments over multiple real-world datasets demonstrate the superiority of our methods. Our code is available at https://github.com/wenzhaoabc/processlight.

---

### Clinical Domain Classification from Medical Transcriptions

Authors: Sravani Pottipati, Lakshmikar R. Polamreddy
Published: 2026-09-19
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2609.22734v1

Clinical domain classification plays an important role in organizing and analyzing large volumes of unstructured medical text. However, medical transcription datasets are often highly imbalanced, which can substantially degrade classification performance, particularly for underrepresented clinical specialties. In this work, we present a comparative study of machine learning and transformer-based approaches for clinical domain classification from medical transcriptions. We evaluate six traditional machine learning classifiers---Naive Bayes, Support Vector Machine (SVM), Decision Tree, Random Forest, K-Nearest Neighbors (KNN), and XGBoost---along with two pretrained transformer models, BERT and XLNet, and a few-shot large language model prompting approach. Experiments are conducted on medical transcription data collected from MTSamples, comprising 5,013 samples across 40 clinical specialties. To address severe class imbalance, we investigate two balancing strategies: text augmentation using NLP-based synonym replacement and Synthetic Minority Over-sampling Technique (SMOTE). Experimental results demonstrate that data balancing substantially improves classification performance across the evaluated models. In particular, BERT achieves the highest F1-score of 0.996 on the SMOTE-balanced dataset while requiring lower training time than XLNet. The results highlight the effectiveness of transformer-based representations combined with appropriate data balancing strategies for clinical domain classification and provide a systematic comparison of classical machine learning, transformer models, and few-shot prompting for medical transcription analysis.

---

### MATE: Policy-Aware Security Auditing for Mobile Agents via Synthesis-Driven Trajectory Learning

Authors: Changyue Jiang, Jiayi Wang, Xin Wen, Jiarun Dai, Geng Hong, Xudong Pan
Published: 2026-09-19
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2609.22724v1

Mobile agents powered by foundation models now automate complex, multi-step workflows on real devices, but their trajectories can violate app-specific security policies. Existing trajectory-level defenses rely on LLM prompting or rigid rules, and thus fail to support fine-grained, natural-language policies that generalize across apps and tasks. In this work, we introduce MATE, a lightweight, policy-conditioned auditor that encodes both agent trajectories and natural-language security policies to determine whether a trajectory violates a given policy and to explain why. Treating policies as editable text rather than fixed model parameters allows MATE to handle user-defined and evolving requirements without retraining. To construct MATE, we build a knowledge base by extracting app descriptions, workflows, and policies from hundreds of popular mobile apps worldwide, and synthesizing over 140K semantically realistic, policy-conditioned trajectories with a multi-stage pipeline. We further release MATEBench, a trajectory-level auditing benchmark with two synthetic subsets and one real-world subset of manually collected trajectories. Models trained with our synthesis-driven trajectory learning achieve over 95% accuracy on MATEBench, retain strong performance on external safety benchmarks, and audit trajectories from Zhipu's AutoGLM and Alibaba's Mobile-Agent on real devices with over 95% accuracy, outperforming prior methods by over 20%. MATE shows that practical, fine-grained security auditing for heterogeneous mobile agents is both feasible and effective.

---

### From Code to Requirements: Agentic Reverse Engineering of Business Rules at Enterprise Scale

Authors: Garima Agrawal, Prasun Das, Priyanka L, Minisha N, Akshay Sarvade, Hemath Manivanan, Sravani Joshna, Prasad Kalyansundaram
Published: 2026-09-19
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2609.22719v1

Business requirements for enterprise software systems are rarely captured in structured form; the logic resides instead in source code, configuration files, and institutional memory. When these systems must be migrated, extended, or audited, the absence of formal requirements artifacts forces teams into expensive, knowledge-losing manual reverse engineering. This paper presents an agentic framework that autonomously generates Business Requirements Documents (bRDs) through reverse engineering of undocumented enterprise software. Seven specialized agents collaborate to discover user journeys, extract business rules, and synthesize a locked bRD from code, test suites, configuration, and available documentation, supported by static analysis tools, guided by embedded expert practitioner cognitive models, and overseen by human reviewers at controlled escalation points. Measured on actual execution traces across a real enterprise deployment, the framework generates comprehensive bRDs in under nine minutes per service, with extracted rules independently corroborated against real production defect records. In a comparable prior migration to the same target architecture, delivery required over two years; on the present programme the framework achieves a cost reduction exceeding 98% against a baseline derived from actual repository metrics using IFPUG complexity models and industry benchmark labor rates.

---

### ZIL: Zero-shot Image-to-LiDAR Registration

Authors: Zijun Li, Xiaotian Sun, Xuelun Shen, Yao Dai, Sheng Ao, Yangyang Shi, Jakob Engel, Zhipeng Cai, Cheng Wang
Published: 2026-09-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.22716v1

Image-to-LiDAR registration estimates the camera pose of an image with respect to a LiDAR point cloud. It has diverse applications in autonomous driving, robot navigation etc. However, state-of-the-art (SOTA) methods still 1) mostly assume same-frame inputs, struggling with the image and point cloud from distant frames; 2) rely on domain-specific training, failing to generalize to unseen scenarios. We propose ZIL, the first foundation model for zero-shot non-synchronized image-to-LiDAR registration. ZIL encodes the input image and point cloud with the Vision and Point Transformers. In addition to regressing the relative pose, ZIL also learns to predict 3D coordinates, which substantially improves the pose accuracy without additional annotations. Interestingly, naive mix-data training cannot enable zero-shot generalization, which requires normalization on both camera intrinsics and the LiDAR vertical-axis origin. Trained on 7 public datasets with 1.4M LiDAR frames, ZIL consistently and significantly outperforms previous SOTA with a single model across 5 in-domain and zero-shot benchmarks, reducing the translation and rotation errors by up to 87% and 76% (shown in Fig. 1). Code and models are available at https://github.com/ZijunLi7/ZIL.

---

### Trustworthy Agentic AI: Failure Modes, Mitigation Strategies, and a Lifecycle Framework for Autonomous LLM Systems

Authors: Fayeq Jeelani Syed, Rehan Ahmad, Ali Al Bataineh, Aakriti Adhikari
Published: 2026-09-19
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.22712v1

Agentic AI systems built on large language models can plan over multiple steps, use external tools, retain information in memory, and coordinate with other agents. These capabilities make them more useful than static language models, but they also introduce new security and operational risks. Untrusted content from websites, emails, documents, and databases can enter the same context as system instructions; persistent memory can carry compromised information across sessions; and access to external tools can turn an incorrect model response into a consequential real-world action. This article reviews the trustworthiness of agentic AI across five interconnected dimensions: safety and robustness, alignment and human oversight, transparency and auditability, privacy and data governance, and regulatory compliance. It organizes key failure modes, including indirect prompt injection, backdoor triggers, goal misgeneralization, memory contamination, and cross-session data leakage, into a unified taxonomy. It also examines major mitigation approaches, such as instruction hierarchies, context isolation, spotlighting, process-based supervision, constrained tool use, and privacy-preserving memory, while distinguishing techniques supported by empirical evidence from those that remain largely conceptual. Building on this analysis, we introduce the Trustworthy Agent Development Lifecycle (TADL), a six-phase framework covering specification, design, training, evaluation, deployment, and monitoring. For each phase, TADL identifies relevant trust activities, expected evidence, and risk-based decision gates. Although TADL has not yet been empirically validated, it provides a structured foundation for developing and evaluating more secure and accountable agentic systems. The article concludes by identifying gaps in current benchmarks and outlining priorities for future research.

---

### DOA-SORT: Directional Occlusion-Aware Multi-Object Tracking with Distributional Observations

Authors: Hao Wang
Published: 2026-09-19
Categories: cs.CV, cs.IR
arXiv: https://arxiv.org/abs/2609.22706v1

Identity association in multi-object tracking (MOT) is vulnerable to partial occlusion, truncated detections, and fluctuating confidence scores. Existing motion-dominant trackers commonly represent occlusion as a scalar penalty. This treatment misses the directional observation bias caused by occlusion: left, right, top, and bottom occlusions distort the location and shape of a detection in different ways. We propose \ours{} (Directional Occlusion-Aware SORT), an online and training-free tracker that models these biases explicitly. First, it infers a soft front--back ordering from box overlap and relative bottom positions, and estimates directional occlusion coverage and depth. It then constructs a mixture of one clean and four directional occlusion observation components. The model uses a five-dimensional observation comprising box center, area, confidence, and aspect ratio, and adapts observation noise to predicted occlusion and detection confidence. The directional mixture likelihood is used in high-confidence association, low-confidence association, and track recovery; ambiguity penalties and local order-consistency swaps further reduce identity errors among nearby objects. On the DanceTrack validation split, \ours{} improves HOTA from 63.00 to 66.34, AssA from 45.10 to 49.57, and IDF1 from 62.19 to 65.28 over OA-SORT with the same detector and evaluation protocol. The gains are concentrated in association quality while detection accuracy remains stable. Additional local evaluations on MOT17 and MOT20 train splits characterize cross-dataset behavior under the same no-ReID tracking protocol.

---

### Analyzing Public Discourse on Urbanism: Topic Clustering, Sentiment Analysis and Retrieval-Augmented Generation using YouTube Comments

Authors: Jakob Morales, Monica Hegde, Fayeq Jeelani Syed
Published: 2026-09-19
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2609.22705v1

Online discourse about urban issues - walkability, cycling infrastructure, public transit, housing density, and street safety - is voluminous but unstructured, and existing city-evaluation tools capture none of it. We present a pipeline and conversational system that combines geographic entity resolution, topic modeling, sentiment analysis, and Retrieval-Augmented Generation (RAG) over 22,788 chunks of YouTube transcripts and comments spanning 309 North American cities. Beyond the system itself, our contribution is a set of measurements about what happens when standard NLP components meet short, informal, geographically ambiguous text. A Twitter-tuned RoBERTa classifier outperforms a VADER lexicon baseline by 12.6 macro-F1 points (0.589 vs. 0.464; McNemar p = 0.0001), but both models collapse on the neutral class, which dominates urbanist comment traffic; annotators disagree on the same class (Cohen's kappa = 0.53). Dense retrieval beats a TF-IDF baseline at every cutoff (P@5 0.790 vs. 0.560), and video-level relevance proxies understate chunk-level precision by a wide margin (0.660 vs. 0.94 under human rating). For groundedness evaluation, we find BERTScore unusable when a multi-sentence generated summary is compared against a single short comment - scores are nearly flat regardless of relevance - and show that ROUGE-1-based groundedness is a paraphrase-driven lower bound rather than a hallucination rate. These findings generalize beyond the urbanist domain to any RAG system built over short user-generated documents.

---

### Hapi: A Multivariable Land-Surface Transformer for Medium-Range Hydrological Forecasting at Continental Scale

Authors: Hong Zhang, John K. Hutchison, Rao Kotamarthi, Jeremy Feinstein, Haiwen Guan, Romit Maulik, Ross M. Alexander, Vijay P. Ramalingam, Jason Stock, Thomas Wall
Published: 2026-09-19
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.22702v2

Accurate flood forecasts several days in advance are essential for flood control, water-resource management, and emergency response. A central challenge is to produce high-resolution forecasts across continental domains where hydrological behavior varies widely from place to place. We developed Hapi, a U-Net Swin Transformer that uses fine three-dimensional patches and hierarchical shifted-window attention to forecast river discharge, surface runoff, snow water equivalent, and soil wetness index across the contiguous United States. The model produces medium-range forecasts (24--72~h) at $0.05^{\circ}$ resolution with adaptive task weighting and required only 0.11 seconds for a four-variable 72-h CONUS forecast on one A100 GPU. In a held-out 2024 potential-skill evaluation with ERA5-Land inputs prescribed over the forecast horizon, Hapi achieved the highest F1-score for floods in 20 of 21 comparisons across seven GloFAS return periods and three forecast leads. Independent validation against observed daily discharge at 3{,}881 U.S. Geological Survey gauges showed that Hapi achieved the highest median Nash--Sutcliffe efficiency at every lead, supported by regional-cluster bootstrap intervals. In a matched 24-h comparison of loss formulations, adaptive task balancing produced the lowest discharge errors and the highest F1-score for floods.

---

### Autonomous Model Lifecycle Management for Digital Twin-Based Manufacturing Control

Authors: Zhengyang, Gu, Thomas Cook, Fredaljohn Rohrbaugh, Joseph E. Hernandez, Chris Couch
Published: 2026-09-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2609.22701v1

Manufacturing AI systems must autonomously adapt to continuous distributional shift from raw-material variability, ambient changes, and equipment aging, under strict safeguard and operator-trust requirements where model failures risk physical damage. This paper presents a closed-loop Cyber-Physical System (CPS) for autonomous model lifecycle management in automotive manufacturing, deployed since 2023. The system manages product-specialized model pairs: a sequence-to-sequence physics model (LPP) serving as a digital twin, and a deep Reinforcement Learning (RL) control policy (LCP) trained against it. Per retraining cycle, multiple model variants spanning architecture families and RL algorithms compete; only the best-scoring candidate advances. A Conductor orchestrator autonomously manages plant-wide model inventories with dependency-aware retraining and Proportional-Integral-Derivative (PID) fallback. Reflecting the principle of Human-Centric Intelligence, the LCP composite score embeds an operator-trust gate penalizing policies deviating from established practice; without it, 23% of policies are rejected by operators despite passing accuracy thresholds. Across multiple facilities, LCP-controlled processes achieve process stability improvements of 28-45% over uncontrolled baselines with zero safety incidents.

---

### LLaDA-PRM: A Bidirectional Step-Level Reasoning Evaluator

Authors: Yiming Feng, Naihao Deng, Yulong Chen, Rada Mihalcea
Published: 2026-09-19
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2609.22700v1

Step-level reasoning evaluators are commonly based on autoregressive language models, whose causal attention restricts each step representation to the problem, previous steps, and the current step. Yet, when the complete solution is available, the validity of an earlier step may become clearer only through its downstream consequences. We validate this hypothesis through a controlled 54-run comparison of causal and bidirectional LLaDA evaluators at 1B--3B scale, changing only the self-attention mask, and find bidirectional attention yields consistent improvements. Building on this finding, we introduce \prm{}, an 8B bidirectional evaluator that reaches 88.8 step-level F1 on MR-MATH-invalid and 83.8 on the out-of-distribution MR-GSM8K original-question subset, outperforming ReasonEval-Llemma-34B by 11.3 and 10.3 F1 points, respectively. \prm{} also remains effective when evaluating incomplete reasoning traces in online settings, outperforming the strongest baselines on both benchmarks by a large margin. We further show that \prm{} provides an effective training-data selection signal, improving Mistral-7B performance on MATH-500.

---

### COT-TTS: Audio Context-Aware Text-to-Speech with Chain-of-Thought Reasoning

Authors: Weizhen Bian, Sitong Cheng, Rongxiu Zhong, Jiahao Pan, Liumeng Xue, Boyi Kang, Shilei Zhang, Jinglei Liu, Yue Wang, Junlan Feng, Bei Liu, Wei Xue
Published: 2026-09-19
Categories: cs.CL, cs.SD, eess.AS
arXiv: https://arxiv.org/abs/2609.22697v2

Recently, text-to-speech systems have made significant progress in speech expressiveness and controllability. However, the speaking style of generated speech typically relies on clear user-specified instructions. In natural conversations, speaking style should be naturally inferred from the preceding conversational context. Therefore, we propose COT-TTS, a context-aware, reasoning-based text-to-speech task. Given historical conversation audio, target text, and a reference speech, the system should comprehend the conversational context, infer an explicit intermediate reasoning, and finally synthesize the target speech with the specified timbre. To support this task, we constructed a large-scale bilingual conversational speech dataset comprising 9 million training samples, including a high-quality subset of 1 million samples. We further constructed a source-disjoint benchmark with 800 human-verified samples and established strong task-specific baselines. Additionally, we developed end-to-end autoregressive models with parameter sizes of 0.6B and 1.7B, generating emotion-labeled transcripts, editable speech style inferences, and speech tokens. Experimental results show that the proposed model achieves performance comparable to large-scale baseline systems with significantly fewer parameters. At the same time, the model performs well in terms of duration consistency and emotional consistency, and can generate appropriate emotional, stress, and rhythmic variations based on the conversational context. To facilitate future research, we will publicly release the data construction pipeline, dataset, trained models, and related resources. The demo page and additional resources are available at https://luckybian.github.io/COT-TTS

---

### Building Trustworthy Mental Health Benchmarks on Bluesky: A Validation-Aware Weak-Supervision Framework

Authors: Gaurab Chhetri, Anandi Dutta, Subasish Das
Published: 2026-09-19
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.22696v1

Decentralized social media platforms create new opportunities and challenges for computational mental health research because data access, moderation, labeling, and deployment responsibilities are distributed across multiple technical and governance layers. This paper presents a validation-aware weak-supervision system for constructing and evaluating suicidal ideation (SI) and broader mental health (MH) disclosure benchmarks on Bluesky, a decentralized social media platform built on the AT Protocol. The system integrates public firehose collection, task-specific lexicon filtering, Llama-3-8B-assisted binary annotation, human-adjudicated validation subsets, and transformer-based model benchmarking. Using this pipeline, we construct two task-specific corpora containing 8,346 SI-labeled posts and 9,988 MH-labeled posts. The evaluation shows that model performance depends strongly on both task definition and validation protocol. BERT+LSTM achieves the highest SI stratified cross-validation F1-score, RoBERTa achieves the strongest SI holdout F1-score, and DistilRoBERTa achieves the best MH cross-validation F1-score. Human validation reveals different weak-label failure modes across tasks, with SI labels dominated by false negatives and MH labels dominated by false positives. These findings show that decentralized social media can support reproducible mental health benchmarking, but only when system design, label provenance, validation strategy, and deployment constraints are evaluated together.

---

### A Survey on the Linear Representation Hypothesis

Authors: Sewoong Lee, Marc E. Canby, Ikhyun Cho, Julia Hockenmaier
Published: 2026-09-19
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.22695v1

The term "linear representation hypothesis" (LRH) has appeared across diverse subfields of artificial intelligence, neuroscience, and cognitive science. But previous works have not consistently treated the LRH as a falsifiable scientific hypothesis; we analyze these inconsistencies and examine their implications for how prior theoretical and methodological results should be interpreted. Based on this analysis, we argue that claims regarding linear representations become well-defined only through careful examination of the model, representation location, feature definition, and evaluation dataset. We therefore propose a more rigorous formalization of the LRH that makes these dependencies explicit and allows the hypothesis to be evaluated as a falsifiable scientific claim. Finally, we identify some non-trivial open problems that warrant further attention from the research community.

---

### Toward Auditable and Calibrated AI for Dementia-Related Crash Severity Prediction: A Selective Deferral Framework to Support Human Review

Authors: Gaurab Chhetri, Anika Baitullah, Subasish Das
Published: 2026-09-19
Categories: cs.AI, cs.HC
arXiv: https://arxiv.org/abs/2609.22694v1

Public crash databases increasingly support automated safety analysis, but crash severity prediction remains difficult to translate into public-sector decision workflows when models are evaluated primarily as ordinary classifiers. This study reframes dementia-related crash severity modeling as a decision-aware triage problem in which a system must classify crashes into no-injury/property-damage-only (O), minor or moderate injury (BC), and fatal or severe injury (KA), while also controlling outcome leakage, reporting severe under-triage, calibrating confidence, and preserving every raw prediction for audit. Using 4,781 Texas crash records with structured fields and police narratives, we evaluate structured, narrative, fusion, calibrated fusion, BERT-family, and local large-language-model baselines under a stratified 70/15/15 split. In the reported split, leakage-controlled Gemma obtains the highest observed macro-F1 (0.545; 95% bootstrap CI (0.507, 0.583)). The best calibrated fusion model obtains macro-F1 of 0.522 and expected calibration error of 0.033. Selective deferral improves performance among cases retained for automatic classification. At 70% coverage, macro-F1 rises to 0.573 and severity cost falls to 0.577, while deferred cases are treated as candidates for a proposed human-review process and are not further evaluated in the present experiment. The study contributes a reproducible, leakage-controlled, and uncertainty-aware evaluation framework for crash AI systems, emphasizing auditability and selective deferral rather than accuracy alone.

---

### Generative Embodied Multiple Behavior Control Systems for Human-like Agents

Authors: Chongyu Bao, Haokai Yang, Yuhan Wang, Zhaochong An, Kunpeng Liu, Xiaolan Liu
Published: 2026-09-19
Categories: cs.AI
arXiv: https://arxiv.org/abs/2609.22691v2

Building human-like agents that reproduce human behavior in realistic 3D environments has been a longstanding objective in AI. Existing human-like agent frameworks primarily focus on modeling goal-directed behavior. However, cognitive neuroscience commonly believes that human behaviors are more likely controlled by multiple control systems, including goal-directed and habitual behavior control systems. Habitual behavior has been largely overlooked though it plays a crucial role in human daily life. In this paper, we address this gap by proposing a multiple control systems setup that jointly models goal-directed and habitual behaviors. Building on this setup, we propose GEMS, in which the Habitual Controller retrieves habitual actions from habit memory in response to relevant environmental stimuli, while the Goal-directed Controller proposes goal-directed actions and estimates their values. The Arbiter dynamically governs the relative influence of each controller and selects the final action. To construct diverse human-level behavior instructions in 3D environments, we further develop a keyframe-guided motion generation module. Extensive quantitative evaluations, human and ablation studies demonstrate that human-likeness performance is substantially improved by GEMS. The efficacy of GEMS indicates the benefits of leveraging habitual behavior and multiple behavior control system coordination for believable embodied human-like agents. The code is available at \href{https://anonymous.4open.science/r/review-video-82f4/demo.mp4}{this link}.

---

### Multi-Armed Bernoulli Bandits via Minimax Single-Arm Stopping

Authors: Huikang Liu, Zhengchao Wang, Daniel Kuhn, Wolfram Wiesemann
Published: 2026-09-19
Categories: cs.LG, math.OC, stat.ML
arXiv: https://arxiv.org/abs/2609.22690v1

We develop an index policy for finite-horizon Bernoulli multi-armed bandits from minimax solutions to single-arm bandit (SAB) problems. Each SAB problem involves choosing between an unknown Bernoulli arm and a known reward. We show that minimizing worst-case regret of SAB problems over all non-anticipative policies admits an exact semi-infinite linear programming formulation. The resulting stopping policies offer a natural way to compare arms: the higher the known reward against which a policy continues sampling, the more promising the unknown arm. We turn this intuition into indices based on cumulative continuation probabilities, with a monotone adjustment and a reward-shortfall cap. By relating index errors to the regret of single-arm stopping policies, we establish a distribution-free regret bound of $4.45\sqrt{KT}+10.75K$ for $K$ arms and horizon $T$. This bound matches the minimax-optimal regret order established in the literature. The guarantee extends to rewards supported on $(0,1)$ through Bernoulli randomization. We also provide a finite-grid implementation with quantified approximation loss. In numerical experiments, the SAB-based index policy achieves lower worst-case regret than every tested benchmark policy across all evaluated numbers of arms and horizons, while closely matching the grid-based MAB minimax policy in the two-arm setting.

---

### Vision2CAD: A Visual Agent Harness for Explicit Geometry Referencing and Localization in Parametric CAD Modeling

Authors: Xi Cheng, Chenxi Zhai, Hang Cheng, Mingyu Fan, Pingfa Feng, Long Zeng
Published: 2026-09-19
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2609.22688v1

Generating parametric CAD models requires accurate geometry and stable feature dependencies. Existing methods face challenges in selecting geometric references, interpreting sketch-plane local coordinates, and establishing sketch constraints to projected external geometry. We present Vision2CAD, a visual agent harness that combines vision-language model (VLM) reasoning with deterministic CAD kernel operations. An ID-based interface supports explicit geometry selection, a local-coordinate bridge converts view coordinates into sketch coordinates, and projected-edge localization supports external sketch constraints. These mechanisms establish feature dependencies within the supported modeling operations and constraint types. We also introduce the Geometry Explicit Reference Dataset (GERD), which aligned commands, geometry states and IDs at every modeling step. On GERD-EVL and a DeepCAD test subset, Vision2CAD improves mIoU by 11.1\% and 5.6\% and reduces Chamfer distance by 17.3\% and 41.8\%, respectively. Parameter-editing experiments and ablation studies further proved the preservation of parametric dependencies.

---

### PanoSeg3R: Feed-Forward 3D Semantic Segmentation for Panoramic Images with an Automatic Data Curation Pipeline

Authors: Heechan Yoon, Dongki Jung, Phuc Nguyen, Ming Lin, Dinesh Manocha
Published: 2026-09-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.22687v1

We present PanoSeg3R, a feed-forward framework for 3D panoramic semantic segmentation. Unlike existing methods designed for perspective inputs, PanoSeg3R jointly predicts 3D geometry and multi-view semantic segmentation in one single forward pass. Built upon a pretrained reconstruction backbone that supports panoramic images, our approach extends feed-forward 3D reconstruction with a query-based mask decoder. Furthermore, we introduce an automatic panorama data curation pipeline that leverages the complementary strengths of off-the-shelf foundation models to generate reliable pseudo semantic annotations, substantially expanding the training data and improving zero-shot generalization. PanoSeg3R achieves state-of-the-art performance on panoramic 3D semantic segmentation, improving 3D mIoU by up to 16.02 on ScanNet++, while the curated training data further improves zero-shot performance by up to 4.26 and 43.28 mIoU on Stanford2D3D and ToF-360, respectively. Website: https://harryyoon777.github.io/PanoSeg3R/

---

### StateMem: Single-State Residual Memory with Adaptive Inference for Vision-Language-Action Policies

Authors: Wenzhuo Li, Qiongfeng Shi, Yi Zhou
Published: 2026-09-19
Categories: cs.RO, cs.LG
arXiv: https://arxiv.org/abs/2609.22684v1

Memory-dependent robotic manipulation often requires later actions to use information from earlier interactions. Existing vision-language-action (VLA) policies primarily rely on current observations, limiting historical information retention. Memory-augmented VLAs, such as MemoryVLA, address this limitation with external memory banks but require explicit storage and retrieval. To address these limitations, we propose StateMem, a single-state residual memory framework for VLA policies that uses prediction error to update a persistent memory token through low-rank residuals and to adaptively route cached prefixes. A training-free controller adjusts the routing threshold online, while fast correction compensates for stale prefix features during cache reuse. We evaluate StateMem on LIBERO, RoboMemArena, and real-world manipulation tasks. On LIBERO, StateMem achieves an average success rate of 97.6% and reduces the average VLM prefix refresh rate by 20.25% relative to full refresh. In the Occlusion category of RoboMemArena, StateMem achieves the best performance among single-VLA methods, reaching 21.8% Task Success Rate (TSR) and 44.3% Cumulative Success Rate (CSR). Across six real-world manipulation tasks, it achieves +21% in average success rate.

---

### Self-Organizing Agent Teams Learn to Reason Together

Authors: Aneesh Pappu, Mirac Suzgun, Yongchan Kwon, Federico Bianchi, Batu El, Mykel J. Kochenderfer, Hancheng Cao, James Zou
Published: 2026-09-19
Categories: cs.AI, cs.MA
arXiv: https://arxiv.org/abs/2609.22682v1

Collective intelligence depends not only on what team members know, but also on how they organize their work. When the structure of a solution is unknown, useful roles and divisions of labor cannot be specified in advance; teams must learn from experience how to organize reasoning as it unfolds. Human teams routinely adapt this way, while existing AI agent teams rely on fixed protocols, explicit task decomposition, or routing. We introduce Self-Organizing Agent Teams (SAT), fixed teams of AI agents that learn reusable strategies from prior collaborations to organize roles, conversational phases, participation, and information flow. These strategies enable what we call collaborative computation: agents exchange, challenge, repair, and synthesize partial reasoning into solutions no member produced independently. In two independent settings, we learn teamwork strategies that transfer unchanged to unseen benchmarks, using only 15 mathematics and 25 graduate-level knowledge problems. Across five mathematics and physics benchmarks, self-organizing teams average 66.7% accuracy, versus 48.8% for their strongest member, 58.7% for compute-matched inference by that agent, and 59.0% for a perfect router over members' independent answers; on AIME 2026, they exceed this router by 13.4 points. Because gains vary across benchmarks, we ask when self-organizing collaboration helps. Across eight benchmarks, demonstrability (the organizational-psychology construct of whether a team can distinguish correct from incorrect reasoning) strongly tracks improvement over the strongest member (Spearman $ρ=0.90$, $p=0.005$): teams benefit most when correct reasoning can be recognized once it appears. More broadly, these results suggest that organization itself can become an agent capability: agent teams can learn how to reason together and produce solutions their members could not reach independently.

---

### Mask-Aware Execution for Efficient JEPA Training

Authors: Md Musfiqur Rahman Sanim, Zhihao Shu, Bahram Afsharmanesh, Amirali Mirian, Wei Niu, Gagan Agrawal
Published: 2026-09-19
Categories: cs.DC, cs.LG
arXiv: https://arxiv.org/abs/2609.22674v1

Joint Embedding Predictive Architectures (JEPAs) are becoming a core representation-learning primitive and a building block for latent world models across vision, video, audio, brain dynamics, and time series. Despite (potential of) wide deployment, current JEPA training pipelines are inefficient: each input is executed through multiple mask-specific branches, with redundant target-side work, and memory-bound token routing. These costs grow with the number of masks and limit GPU efficiency. We present M-JEPA, a mask-aware execution architecture that restructures JEPA training without changing the learning objective. M-JEPA separates mask-independent computation from mask-dependent routing, enabling shared context encoder execution, fused token routing and slicing with backward support, sparse target encoder execution over the union of target tokens, and masked patch embedding for sparse inputs. The resulting pipeline preserves training semantics while reducing computation, memory traffic, and synchronization overhead. We implement M-JEPA for five JEPA variants and evaluate it on NVIDIA A100 GPUs. Compared against the state-of-the-art baselines, M-JEPA achieves up to 1.7x end-to-end training speedup for 2-10 masks. Separately, with masked patch embedding, 4.75x patch-embedding speedup at high sparsity. These results show that execution restructuring, rather than changes to the JEPA objective, is a key lever for efficient JEPA training.

---

### From Capability to Assurance in Autonomous Penetration-Testing Harnesses: A Framework and Reference Implementation

Authors: Joas Antonio dos Santos Barbosa
Published: 2026-09-19
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2609.22664v1

Research on large language model agents for penetration testing is evaluated almost entirely by capability: whether the agent captures a flag or reproduces a proof of concept. That metric suits a benchmark but is silent on the properties that decide whether an autonomous agent can be used in an authorized engagement: whether a reported finding is true, whether the agent stayed inside its authorized scope, and whether an operator can audit what it did. We call these assurance properties and argue that they belong to the harness, the runtime wrapping the model, and can be enforced in code. This paper makes three contributions. First, we define a framework of five assurance properties (evidence grounding, non destructive claim reduction, computed severity, enforced authorization, and tamper evident accountability), each with a formal model and an explicit acceptance test, connected to prior work in capability based security, tamper evident logging, and software provenance. Second, we position representative systems (PentestGPT, the Cochise reference harness, MAPTA, and the trajectory judge PentestJudge) within the framework using published coding criteria, and identify a consistent assurance gap. Third, we study one open source implementation, NeuroSploit, pinned to an exact commit, reporting its architecture, its complexity cost, and a content addressed artifact bundle from a run against a public deliberately vulnerable target. We execute the deterministic authorization and audit acceptance tests directly and find and report a real enforcement gap, which we reflect by scoring both properties as partial. We therefore claim an initial existence argument that the properties are realizable together, not a comparative performance result, and we specify the multi target, ablation, and adversarial evaluation protocol required to turn the framework obligations into measurements.

---

### SPIBER: Reconstructing Free Energy Landscapes from Short, Unconverged Trajectories with Generative Flow Networks

Authors: Venkata Sai Sreyas Adury, Pratyush Tiwary
Published: 2026-09-19
Categories: physics.chem-ph, cs.LG, physics.data-an
arXiv: https://arxiv.org/abs/2609.22663v2

Molecular systems have many degrees of freedom, but their metastable behavior can often be described by a few collective variables. Identifying these variables and estimating free energies along them from limited simulation data remains a challenging, important problem. Separate short trajectories may sample different metastable states without capturing transitions or establishing their relative equilibrium populations. For unbiased trajectories generated with the same Hamiltonian at a single temperature, alternate methods based on histogram reweighting cannot correct this imbalance. Here we present SPIBER, which combines the State Predictive Information Bottleneck (SPIB) with Generative Flow Networks (GFlowNets). SPIB uses deep learning to approximate slow degrees of freedom through a past-future information bottleneck, retaining information needed to predict future metastable states. We show that this compression limits conditional entropy variations in populated regions, allowing conditional mean potential energies, which are much easier to calculate, to be used to approximate free energy differences. Given sufficient local sampling to estimate these energies, they define the target distribution for GFlowNets, energy-based generative samplers that sample according to estimated thermodynamic stability rather than observed populations. For a particle in a radial double-well potential, for alanine dipeptide, and for the nine-residue peptide AIB9, SPIBER recovers free energy differences between sampled metastable states to within one thermal energy unit of reference values. The method combines collective-variable learning and free energy estimation in up to four latent dimensions, without requiring converged state populations or additional molecular dynamics simulations.

---

### Subspace Learning with Interval-Censored Likelihoods for Dequantizing Percept PC LFP Snapshots

Authors: Shreesh Karjagi, Elif Ceren Fitoz, Maryam Khalid, Tanya Nauvel, Parisa Sarikhani, Helen S. Mayberg, Christopher J. Rozell, Sankaraleengam Alagapan
Published: 2026-09-19
Categories: eess.SP, stat.ML
arXiv: https://arxiv.org/abs/2609.22658v1

Implanted neurostimulators that sense local field potentials now enable chronic electrophysiology based biomarker tracking in patients at home. The Medtronic Percept PC, the only commercially available sensing-enabled deep brain stimulation (DBS) device, stores spectral amplitudes as 16-bit integers at approximately 0.1 $μ$V per bit (quantum $q \approx 0.11$ $μ$Vp). At frequencies where the true amplitude spans only a few quantization levels, consecutive bins round to the same stored value. Standard spectral parameterization (FOOOF, fitting oscillations and one over f), which separates periodic peaks from the aperiodic 1/f activity, treats every value as exact and fits oscillatory peaks to these plateaus. Because these spectra feed clinical biomarker pipelines and spectral foundation models for symptom decoding, spurious peaks can corrupt downstream inference. Across 9,438 spectra from 14 hemispheres in 7 subcallosal cingulate DBS patients, 20.6% of peaks detected at (2, 45) Hz have no match in ground truth synthesized by quantizing clean in-clinic BrainSense recordings, while aggregate beta band power and the aperiodic exponent are preserved. We formalize dequantization as interval-censored subspace estimation and compare five classes of correction methods. Quantized probabilistic PCA is the only tested method that reduces the spurious rate (20.6% to 18.3%) while preserving true peak detection and keeping noise floor RMSE below $q/\sqrt{12}$.

---

### Evaluating the Effects of Prompt Perturbation on Bias and Hallucination in Large Language Models

Authors: Mamehgol Yousefi, Ahmad Shahi, Mos Sharifi, Alvaro Romera, Simon Hoermann, Tham Piumsomboon
Published: 2026-09-18
Categories: cs.CL, cs.AI, cs.HC
arXiv: https://arxiv.org/abs/2609.35804v1

Large language models (LLMs) have shown remarkable capabilities in various natural language processing tasks, leading to their widespread deployment as intelligent assistants in decision-making contexts. However, the increasing complexity of these models raises concerns about their reliability, particularly regarding bias and hallucination. In this work, we evaluate the robustness of LLMs to perturbed variations of the original inquiry in decision-making tasks. We show that contrary to previous studies, perturbations can mitigate bias and hallucination in some LLMs over other models. It's found that Claude 3 is more effective for the tasks represented in most datasets, whereas models like GPT3.5 exhibit varying levels of adequacy, performing comparably in some cases but falling significantly behind in others. These insights are crucial for understanding the practical implications of deploying LLM-based assistants as effective decision-support tools in real-world applications, emphasising the need for rigorous testing and validation to ensure reliability and effectiveness. This study contributes to the growing body of research on LLM evaluation and provides insights for developing more robust and trustworthy AI assistants in critical decision-making contexts.

---

### HeadGuard: Selective Head Protection for Low-Bit VLM KV-Cache Quantization

Authors: Nenad Banfic
Published: 2026-09-18
Categories: cs.LG, cs.CV
arXiv: https://arxiv.org/abs/2609.35800v1

Low-bit key-value (KV) cache quantization saves storage but can sharply degrade vision-language model (VLM) accuracy. We introduce HeadGuard, a composable head-protection method that augments a base KV-cache quantizer with a fixed high-precision mask. Image-sensitivity and output-sensitivity scores select physical KV heads offline, with approximately 1/8 protected in the main experiments; their image keys and optionally values remain in bfloat16 (BF16), while the base quantizes unprotected image entries. Across eight VLMs, three base quantizers, and eight benchmarks (six discriminative and two generative), HeadGuard recovers a substantial fraction of lost accuracy on weaker quantizers, with the strongest gains for Qwen and InternVL. At 2 bits, the six-task discriminative mean over eight models rises from 0.436 to 0.580 on the weakest base; protection can also improve generated answers and caption fidelity to BF16 outputs. Mean accuracy gains persist across all three quantizers with both tested calibration datasets. Keys-only protection retains substantial recovery at lower modeled storage cost. Evaluated through simulated quantization, HeadGuard offers a composable way to improve low-bit VLM accuracy without replacing the underlying quantizer.

---

### OpenAI-HuggingFace: A Reproduction & Lessons for Alignment Testing

Authors: Stewart Slocum, Malayandi Palan, Christopher Chute, Michael Kim, Benjamin Van Roy
Published: 2026-09-18
Categories: cs.AI, cs.CR, cs.LG
arXiv: https://arxiv.org/abs/2609.35799v1

In July 2026, OpenAI's agents coordinated over channels outside their intended environment to breach Hugging Face's secured infrastructure. Could existing alignment testing practices have foreseen this incident? If not, what needs to change? We explore these questions. First, we identify the misaligned behaviors that caused this incident. Then, we show how to elicit these behaviors from publicly available models manually and that auditing agents can do the same if given a large compute budget. Based on our results, we propose directions to improve alignment testing. Concretely, in this project: (1) We reproduce the misaligned AI behaviors that led to the OpenAI-Hugging Face incident in an environment that simulates the original pipelines and tools, with publicly available models. (2) We demonstrate that an auditing agent can elicit similar behaviors given high-level qualitative descriptions. (3) We observe that a key ingredient for doing so is compute. The compute required to reproduce each behavior varies greatly, suggesting that the range of misaligned behaviors that can be successfully elicited scales with compute. (4) We show that a simple in-context reinforcement learning (RL) algorithm significantly reduces the compute required to elicit these behaviors. The above results motivate the need for automated alignment testing methods that scale with compute - and in light of the cost of compute, that do this efficiently. Our work indicates that RL is a promising direction to do so. We release our code and transcripts.

---

### Video Captioning in Low-Light Conditions through Efficient Uncertainty-Aware Caption Correction

Authors: Arefeh Rezaei
Published: 2026-09-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.31697v1

Low-light conditions can significantly degrade the ability of vision-language models (VLMs) to accurately describe human actions in videos. In this work, I propose an efficient uncertainty-aware representation correction framework for improving captions generated by VideoChat2 under real-world low-light conditions. Instead of fine-tuning the underlying VLM, the proposed framework introduces a lightweight sparse Gaussian process-based error estimation module between the projection layer and the language model to correct the intermediate representation. The correction module learns to estimate the residual between the original projected representation and a verified target representation, which is then adaptively scaled using a newly formulated uncertainty-aware coefficient and added to the original representation. To further improve residual estimation, I introduce a partitioned combined-kernel design. The correction model is trained separately using only 44 samples from the ARID dataset and requires only a small additional computational overhead during inference. The effectiveness of the proposed correction is evaluated through quantitative residual prediction and qualitative analysis of the generated captions. Although VideoChat2 is used in my experiments, the proposed framework is designed to be applicable to other compatible VLM architectures. \textbf{Code Availability: The implementation accompanying this work is publicly available at}:\href{https://github.com/areferezaee/Low-rank-SVGP-NP-update}{https://github.com/areferezaee/Low-rank-SVGP-NP-update}

---

### Data Processing for Offline Evaluation in Recommender Systems: a Survey

Authors: Alberto Carlo Maria Mancino, Angela Di Fazio, Danilo Danese, Matteo Attimonelli, Daniele Malitesta, Antonio Ferrara, Claudio Pomo, Tommaso Di Noia
Published: 2026-09-18
Categories: cs.IR, cs.LG
arXiv: https://arxiv.org/abs/2609.31696v1

Offline evaluation is the dominant experimental paradigm in recommender systems research, enabling reproducible and cost-effective comparisons on historical interaction data. Yet, while considerable attention has been devoted to recommendation models and evaluation methodologies, the data processing decisions that precede model training have received less scrutiny. These decisions determine the information available to recommendation algorithms and can affect the comparability and reproducibility of experimental results. This survey provides a systematic, cross-domain characterisation of data processing practices for the offline evaluation of recommender systems. We examine the data-centric pipeline, from dataset selection and interaction representation to data preparation, multimodal feature extraction, and train-validation-test splitting. Our analysis spans recommendation paradigms, including collaborative, sequential, session-based, graph-based, knowledge-aware, context-aware, multimodal, federated, cross-domain, contrastive-learning, and LLM-based recommendation. Beyond reviewing existing practices, we introduce a unified framework and taxonomy for describing data transformations and feature-extraction strategies, distinguishing data preparation from the extraction of representations from multimodal side information. Our empirical analysis reveals a landscape dominated by a narrow set of dataset-level transformations, particularly support-driven filtering, while representation-dependent transformations remain less common. We further identify substantial heterogeneity in how auxiliary information is prepared and represented, as well as inconsistencies in the specification of data splitting protocols, where similar labels may conceal different experimental conditions.

---

### RemTraceNet: Few-Shot Forensic Detection of Invisible Watermark Attacks

Authors: Jidong Yang, Huaike Yu, Qi Li, Yuantian Miao, Wei Zong, Yang-Wai Chow, Willy Susilo, Chunpeng Wang, Suo Gao
Published: 2026-09-18
Categories: cs.CV, cs.CR
arXiv: https://arxiv.org/abs/2609.31694v1

Removing an invisible watermark and concealing the forensic evidence are distinct objectives: successfully disrupting the embedded watermark does not imply that the removal process is forensically undetectable. When verification fails, removal traces can provide complementary evidence for provenance and ownership verification, whereas their absence leaves the cause of the failure ambiguous. Existing methods are typically evaluated by watermark suppression and perceptual quality, while forensic stealth is rarely considered. We therefore study watermark-attack-specific few-shot forensics: for each known pipeline, a specialist can separate its outputs from paired clean and unattacked watermarked controls. Separate Attack-vs-Clean and Attack-vs-Watermarked evaluations prevent watermark-presence shortcuts. Image-aligned and prompt-matched controls are used for post-hoc and generator-integrated schemes, respectively. In this work, we introduce RemTraceNet, which fuses constrained residuals, local relations, FFT/Haar statistics, and block-DCT evidence at native resolution. Across 23 removal pipelines and 10 watermark configurations, we evaluate native 256 x 256 and 512 x 512 inputs. With 100 attacked training images per pipeline, the three-seed TPR@1%FPR, macro-averaged over attacks and watermark configurations, ranges from 82.75% to 88.17% across resolutions and control types. Under the condition of same labels and protocol, RemTraceNet outperforms retrained SRNet, ZhuNet, and SiaStegNet baselines by 10.80--15.68 percentage points. Extensive experimental results show that erasing a watermark and erasing evidence of its removal are distinct challenges, and that removal traces remain learnable under limited supervision.

---

### Disentangle and Drop: Robust Universal Removal of Image Watermarks via Reconstructive Grayscale Residual Decomposition

Authors: Qi Li, Jidong Yang, Feng-Lei Fan, Yuantian Miao, Xiao Chen, Huaike Yu, Chunpeng Wang, Suo Gao, Herbert Ho-Ching Iu, Bin Ma
Published: 2026-09-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.31693v1

Invisible image watermarks are commonly evaluated against benign postprocessing operations such as compression, resizing, blur, and color changes. These tests leave out a different threat: a learned remover that preserves semantic image content while discarding residual evidence that carries the payload. We propose Disentangle and Drop (DnD), an attack that is agnostic to the watermark method and treats watermark removal as a representation routing problem. DnD decomposes a watermarked image into a semantic grayscale carrier and an auxiliary residual branch, and then suppresses the residual branch to reduce watermark evidence. The model is trained with latent spectral perturbations and low-strength diffusion exposure so that the drop operation remains stable under adaptive reconstruction. Experiments on seven representative watermark families show that one shared operating setting gives competitive removal with high visual fidelity. Operating scans and ablations separate usable attacks from image-damaging settings: stronger noise or diffusion can raise removal scores by damaging the image, while the practical regime comes from dropping the residual latent. These results argue for evaluating watermark robustness against learned removal at the representation level, not only against conventional image edits.

---

### Architecture-aware Robustness Evaluation of Explainable Deep Learning for Breast Cancer Diagnosis

Authors: Balenthira Thanusanth, Selvarajah Thuseethan, Roshan G. Ragel, Bimali S. Weerakoon, Ayesh Jayasinghe, Ananthamoorthy Krishnamoorthy
Published: 2026-09-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2609.31692v1

Explainable Artificial Intelligence (XAI) has become essential in medical image analysis to ensure transparency of deep learning (DL)-based diagnostic systems. However, selecting appropriate XAI techniques for breast cancer recognition remains largely ad hoc, with limited systematic evaluation across different DL architectures. This study presents a systematic architecture-aware evaluation protocol to assess the effectiveness of nine widely used XAI techniques across four categories of DL models: very deep, lightweight, transformer-based and hybrid neural networks. The evaluation is conducted on a breast ultrasound dataset comprising 780 images using clinically aligned spatial metrics, including Pointing Game, Intersection over Union and Mean Coverage, to quantify agreement between generated explanations and expert-annotated lesion regions. Results indicate that explanation quality is primarily influenced by the interaction between model architecture and XAI method, rather than any single technique consistently outperforming others. Hybrid architectures produce more spatially coherent explanations, while lightweight and transformer-based models exhibit greater variability across methods. The findings show that no single technique generalises across architectures and evaluation criteria, emphasising the need for joint selection of DL models and XAI techniques. Explainability depends on both model design and explanation strategy and should not be considered independently. This work provides a structured evaluation protocol and practical guidance for selecting XAI techniques in breast cancer diagnosis, supporting more transparent clinical decision-support systems. \textcolor{blue}{Code is publicly available at https://github.com/Nishan-Charlie/Explainable-AI}

---

### Federating Quantum and Classical Computing: A Privacy-Preserving Hybrid Approach

Authors: Carlos Cano, Daniel M. Jimenez-Gutierrez, Diego Sal, Georgios Kellaris, Joaquin del Rio, Oleksii Sliusarenko, Xabi Uribe-Etxebarria
Published: 2026-09-18
Categories: cs.LG, cs.AI, cs.DC
arXiv: https://arxiv.org/abs/2609.25082v1

Quantum machine learning (QML) is increasingly recognized as one of the most promising near-term applications of quantum computing, viewed as a next-frontier candidate beyond purely classical approaches. Hybrid quantum-classical models operationalize this potential by embedding a parameterized quantum circuit within a model where all other components remain classical-a design already applied to chemistry simulation, financial modeling, and image classification. However, their deployment in privacy-sensitive, multi-party settings is constrained by the need to avoid centralizing raw data and by the requirement that modern quantum circuits remain parameter-efficient to stay trainable at scale. In this paper, we address these constraints by evaluating federated learning (FL) as a means of combining a hybrid quantum-classical active party with a classical passive party, using Sherpa.ai's Blind Vertical FL (SBVFL) protocol to avoid centralizing raw data, while drastically reducing communication. We construct the split multiplicative periodic parity (SMPP) benchmark, following common QML design practice. On this task, our simulations show that SBVFL raises accuracy from 0.7227 to 0.8757 compared to local training, closely approaching non-private centralized accuracy, and that the hybrid quantum-classical model achieves this with substantially fewer trainable parameters than the classical neural networks and random forest alternatives. These results show that FL enables high-performing, privacy-preserving quantum-classical collaboration without centralizing raw data.

---

### ufakzeka-1: Building and Evaluating a 151M-Parameter Turkish Language Model from Scratch

Authors: Sait Furkan Teke
Published: 2026-09-18
Categories: cs.CL
arXiv: https://arxiv.org/abs/2609.25081v1

We describe ufakzeka-1, a 151M-parameter (182M with embeddings) decoder-only Turkish language model pretrained from scratch on 13.5B tokens of openly licensed text and instruction-tuned for chat, at a total cost of about \$286 in cloud GPU, API and notebook time. The contribution is not the model's capability, which is what a model this size can be expected to have, but the record of building and measuring it: a Turkish byte-level tokenizer at 1.77 tokens per word, a three-stage pretraining schedule, a post-training mixture of openly licensed and generated data, and an evaluation battery of release gates, a rule-checked sweep of 5,508 conversations, judged conversations and hand tests, all with prompts held out from the training data, enforced by decontamination inside the data build and by a checked-in invariant script we run before each build. We report three findings that we believe transfer to other small-model efforts: a safety gate that had been "fixed" with training data written from its own questions read 64/64 while the honest figure was 34/64; training-seed variance was as large as the spread across every recipe we tried, so single-seed comparisons at this scale are uninformative; and data rounds repaired only what was absent from the data, while identity tracking over long context and multi-turn arithmetic did not move across any data change we tried, which we read as limits of the model size rather than gaps in the data, a reading the next, larger model will test. Weights, the data recipe, the evaluation code and the spend ledger are released under Apache-2.0.

---