# Research Papers — 2026-04 · bundle page 19 of 233

> Full-text markdown of papers #901–950 of 11612 published in 2026-04, newest first. AI-readable mirrors of the full 11612-paper 2026-04 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-04/page/19
[← Previous bundle](https://aipod.dev/papers/archive/2026-04/page/18.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-04/page/20.md)

### Adaptive Confidence Intervals in Efron's Gaussian Two-Groups Model

Authors: Qiaosen Wang, Shuwen Chai, Chao Gao
Published: 2026-04-28
Categories: math.ST, stat.ME, stat.ML
arXiv: https://arxiv.org/abs/2604.26992v2

Robust uncertainty quantification is increasingly important in modern data analysis and is often formalized under Huber's model, which allows an $\varepsilon$-fraction of arbitrary corruptions. In many experimental sciences, however, the measurement protocol is well controlled, and contamination is more plausibly introduced upstream. Motivated by this noise-oblivious nature of adversaries, we study confidence intervals for the null location parameter $θ$ in Efron's Gaussian two-groups model, where an unknown fraction $\varepsilon$ of observations have arbitrarily shifted means, but all samples share the same law of additive Gaussian measurement noise with variance $σ^2$. We characterize the minimax-optimal length among confidence intervals with a prescribed coverage level uniformly over the unknown contamination proportion and all noise-oblivious adversaries. Although prior work has shown that the minimax point estimation rate of theta does not deteriorate when $\varepsilon$ becomes unknown, our results reveal that, with a given $σ^2$, the minimax-optimal length of confidence intervals that are adaptive to unknown $\varepsilon$ is of order $σ(n^{-1/4}+\varepsilon^{1/2}/\max\{1, \log(en \varepsilon^2)\}^{1/2})$, which is polynomially worse than the optimal length when $\varepsilon$ is known. When the variance $σ^2$ is also unknown, we show a further degradation: no adaptive confidence interval can be shorter than $Ω(σn^{-1/8})$. Algorithmically, we introduce a Fourier-based certification procedure built on Carathéodory's positive-semidefiniteness constraints. By scanning candidate points and accepting those whose residual characteristic function is certifiably consistent with a Gaussian location mixture, our algorithm attains the minimax lower bound in the known-variance setting and is computable in polynomial time.

---

### People-Centred Medical Image Analysis via Fairness-Aware Human-AI Cooperation

Authors: Zheng Zhang, Milad Masroor, Cuong Nguyen, Tahir Hassan, Yuanhong Chen, David Rosewarne, Kevin Wells, Thanh-Toan Do, Gustavo Carneiro
Published: 2026-04-28
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2604.26991v2

Machine learning models for medical image analysis often exhibit subgroup-dependent performance, which impacts how decisions should be allocated between automated systems and human experts under limited resources. Prior work on AI fairness and human-AI cooperation, including learning to defer (L2D) and learning to complement (L2C), typically addresses these problems in isolation. We propose People-Centred Medical Image Analysis (PecMan), a framework for fairness-aware human-AI co-operative classification that jointly models subgroup-dependent reliability, decision allocation, and collaborative prediction. PecMan combines subgroup-specialised predictors with a gating and consolidation mechanism that dynamically assigns cases to automated models, human experts, or their combination, without requiring sensitive attributes at test time. We also introduce the FairHAI benchmark for evaluating trade-offs between predictive accuracy, subgroup equity, and human involvement. In addition, we provide a theoretical analysis of multi-agent gating via selection regret and characterise fairness-coverage trade-offs under input-dependent allocation. Experiments across multiple medical imaging datasets demonstrate that PecMan achieves consistently improved trade-offs compared to methods that address fairness or human-AI cooperation separately.

---

### BatteryPass-12K: The First Dataset for the Novel Digital Battery Passport Conformance Task

Authors: Tosin Adewumi, Martin Karlsson, Lama Alkhaled, Marcus Liwicki
Published: 2026-04-28
Categories: cs.CL
arXiv: https://arxiv.org/abs/2604.26986v1

We introduce a novel task of digital battery passport (DBP) conformance classification and introduce the first public benchmark for the task: BatteryPass-12K, created synthetically from real pilot samples. This is as the EU's battery regulation on DBPs comes into effect soon and there exists no public dataset. We evaluated 22 language models (LMs) in zero-shot inference, spanning small LMs (SLMs), mixture of experts (MoEs), and dense LLMs. We also conducted analysis, additional evaluations of few-shot inference and prompt-injection attacks to find that (1) Thinking models have the best performance (with GPT-5.4 scoring 0.98 (0.03) and 0.71 (0.22) on average as F1 (and confidence interval at 95%) on the validation and test sets, respectively), (2) few-shot examples improve performance significantly, (3) generally capable frontier models find the task challenging, (4) merely scaling model parameters does not necessarily lead to improved performance, as SLMs outperformed some LLMs, and (5) prompt-injection attacks degrade performance. We note that BatteryPass-12K, though limited to real pilot samples, may be useful for other known or emerging tasks in the battery domain, e.g. lifecycle reasoning. We publicly release the dataset under a permissive licence (CC-BY-4.0).

---

### Simple Self-Conditioning Adaptation for Masked Diffusion Models

Authors: Michael Cardei, Huu Binh Ta, Ferdinando Fioretto
Published: 2026-04-28
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2604.26985v2

Masked diffusion models (MDMs) generate discrete sequences by iterative denoising under an absorbing masking process. In standard masked diffusion, if a token remains masked after a reverse update, the model discards its clean-state prediction for that position. Thus, still-masked positions must be repeatedly inferred from the mask token alone. This design choice limits cross-step refinement. To address this limitation, this paper proposes a simple, yet effective, post-training adaptation for MDMs that conditions each denoising step on the model's own previous clean-state predictions. The resulting method, called Self-Conditioned Masked Diffusion Models (SCMDM), requires minimal architectural change, does not introduce a recurrent latent-state pathway, does not rely on an auxiliary reference model, and adds no extra denoiser evaluations during sampling. This is an important departure from partial self-conditioning approaches which requires expensive model training from scratch. In particular, the paper shows that partial self-conditioning, including the commonly used 50% dropout strategy for training self-conditioned models from scratch, is suboptimal in the post-training regime. Instead, once the model's self-generated clean-state estimates become informative, the specialization to refinement is preferable to mixing conditional and unconditional objectives. SCMDM is evaluated across multiple domains, demonstrating consistent improvement over vanilla MDM baselines, achieving nearly a 50% reduction in generative perplexity on OWT-trained models (42.89 to 23.72), alongside strong improvements in discretized image synthesis quality, small molecular generation, and enhanced fidelity in genomic distribution modeling.

---

### Monitoring Neural Training with Topology: A Footprint-Predictable Collapse Index

Authors: Alexander Kalinowski
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.26984v1

Representational collapse, where embeddings become anisotropic and lose multi-scale structure, can erode downstream performance long before performance metrics react. We propose an online, topology-aware monitor for evolving neural representations that couples Modular Morse Homology Maintenance (MMHM) with a composite Collapse Index (CI). Instead of rebuilding complexes each epoch, we apply sparse edits at a fixed scale and maintain a discrete Morse matching, yielding fast, incremental updates. Across LLM fine-tuning and temporal KGE training, CI provides a low-latency early-warning signal suitable for in-training interventions. Code and experimental scripts will be released publicly

---

### Value-Aware Product Recommendation by Customer Segmentation using a suitable High-Dimensional Similarity Measure

Authors: María Florencia Acosta, Rodrigo García Arancibia, Pamela Llop, Mariel Lovatto, Lucas Mansilla
Published: 2026-04-28
Categories: cs.IR, cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2604.26983v1

This paper presents a novel value-aware approach to product recommendation that simultaneously addresses the high dimensionality and sparsity of user-item data while explicitly incorporating the contribution of each product and user to overall sales revenue. The proposed framework encodes revenue contributions in the user-item matrix and computes customer similarity directly on this basis using suitable distance measures. This enables the segmentation of users according to the revenue-based similarity of their purchase baskets and supports recommendations aligned with profitability objectives. We compare conventional similarity metrics with a novel alternative tailored to high-dimensional contexts and propose three recommendation strategies based on revenue share, product popularity, and expected profit generation. The effectiveness of the proposed method is validated through simulation experiments and a real-world application using the UCI Online Retail dataset.

---

### Budget-Constrained Online Retrieval-Augmented Generation: The Chunk-as-a-Service Model

Authors: Shawqi Al-Maliki, Ammar Gharaibeh, Mohamed Rahouti, Mohammad Ruhul Amin, Mohamed Abdallah, Junaid Qadir, Ala Al-Fuqaha
Published: 2026-04-28
Categories: cs.IR, cs.LG
arXiv: https://arxiv.org/abs/2604.26981v1

Large Language Models (LLMs) have revolutionized the field of natural language processing. However, they exhibit some limitations, including a lack of reliability and transparency: they may hallucinate and fail to provide sources that support the generated output. Retrieval-Augmented Generation (RAG) was introduced to address such limitations in LLMs. One popular implementation, RAG-as-a-Service (RaaS), has shortcomings that hinder its adoption and accessibility. For instance, RaaS pricing is based on the number of submitted prompts, without considering whether the prompts are enriched by relevant chunks, i.e., text segments retrieved from a vector database, or the quality of the utilized chunks (i.e., their degree of relevance). This results in an opaque and less cost-effective payment model. We propose Chunk-as-a-Service (CaaS) as a transparent and cost-effective alternative. CaaS includes two variants: Open-Budget CaaS (OB-CaaS) and Limited-Budget CaaS (LB-CaaS), which is enabled by our ``Utility-Cost Online Selection Algorithm (UCOSA)''. UCOSA further extends the cost-effectiveness and the accessibility of the OB-CaaS variant by enriching, in an online manner, a subset of the submitted prompts based on budget constraints and utility-cost tradeoff. Our experiments demonstrate the efficacy of the proposed UCOSA compared to both offline and relevance-greedy selection baselines. In terms of the performance metric-the number of enriched prompts (NEP) multiplied by the Average Relevance (AR)-UCOSA outperforms random selection by approximately 52% and achieves around 75% of the performance of offline selection methods. Additionally, in terms of budget utilization, LB-CaaS and OB-CaaS achieve higher performance-to-budget ratios of 140% and 86%, respectively, compared to RaaS, indicating their superior efficiency.

---

### Multibit neural inference in a N-ary crossbar architecture

Authors: Anatole Moureaux, Anthony Lopes Temporao, Flavio Abreu Araujo
Published: 2026-04-28
Categories: cs.AR, cs.AI, cs.ET
arXiv: https://arxiv.org/abs/2604.26979v2

In-memory computing (IMC) is a paradigm that enables neural network inference by computing analog matrix-vector multiplications (MVM) directly in memory crossbar arrays, with the potential for energy efficiency gains over conventional von Neumann architectures. In this work we present a simulation framework for N-ary crossbar architectures that retrieves MVM results with minimal implementation assumptions. The XOR and MNIST classification tasks were successfully inferred using a simulated crossbar array of (4x4) 4-states magnetic tunnel junctions (MTJ). MNIST accuracy reached 93.56% (vs. 97.56% software baseline). PCA dimensionality reduction was shown to drastically lower the number of required operations and improve the software baseline, for only a modest reduction in crossbar inference accuracy. We identified weight quantization as the primary error source, and studied its impact alongside systematic non-idealities and random noise. We find that cell-specific random noise is less detrimental than systematic errors due to averaging across the array. Finally, we demonstrate an optimal number of states per cell that balances quantization error against resistance state resolution to minimize total MVM error.

---

### Defeasible Conditional Obligation in a Two-tiered Preference-based Semantics (Extended Version)

Authors: Xavier Parent
Published: 2026-04-28
Categories: cs.LO, cs.AI
arXiv: https://arxiv.org/abs/2604.26977v3

In response to a concern raised by Horty, this paper develops a two-tiered, preference-based semantic framework for modeling defeasible conditional obligations. The paper extends a Hansson-Lewis style preference semantics for dyadic deontic logic by incorporating a nonmonotonic reasoning mechanism that enables previously derived obligations to be withdrawn when new, potentially conflicting information comes in. The account is bi-preferential: two orderings--ideality and normality--on worlds are employed to address shortcomings in earlier approaches, with a separate ranking method for each. At the nonmonotonic layer, a number of postulates are considered, including antecedent strengthening, inclusion and no-drowning. A connection is established with so-called constrained input/output (I/O) logic--an existing standard for normative reasoning based on a different methodology.

---

### Fitting Horn DL Ontologies to ABox and Query Examples: A Tale of Simulation Quantifiers and Finite Models

Authors: Marvin Grosser, Carsten Lutz
Published: 2026-04-28
Categories: cs.LO, cs.AI
arXiv: https://arxiv.org/abs/2604.26976v2

We study the problem of fitting a description logic (DL) ontology to a given set of positive and negative examples that take the form of an ABox and a Boolean query. While previous work has investigated this problem for the expressive DLs ALC and ALCI, we here focus on the Horn DLs EL and ELI, as well as their extensions with the bottom concept. As the query language, we consider atomic queries (AQs), conjunctive queries (CQs), and unions thereof (UCQs). We provide characterization of the existence of a fitting ontology based on simulations, use them to develop decision procedures, and clarify the exact computational complexity. For AQs, the problem is in PTime for both EL and ELI. For CQs and UCQs, it is $Σ_2^P$-complete for EL and ExpTime-complete for ELI. Adding the bottom concept does not change any of these complexities. Interestingly, moving from ALC and ALCI to EL and ELI introduces additional technical challenges rather than simplifying the matter.

---

### Lifting Embodied World Models for Planning and Control

Authors: Alex N. Wang, Trevor Darrell, Pavel Izmailov, Yutong Bai, Amir Bar
Published: 2026-04-28
Categories: cs.CV, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2604.26182v2

World models of embodied agents predict future observations conditioned on an action taken by the agent. For complex embodiments, action spaces are high-dimensional and difficult to specify: for example, precisely controlling a human agent requires specifying the motion of each joint. This makes the world model hard to control and expensive to plan with as search-based methods like CEM scale poorly with action dimensionality. To address this issue, we train a lightweight policy that maps high-level actions to sequences of low-level joint actions. Composing this policy with the frozen world model produces a lifted world model that predicts a sequence of future observations from a single high-level action. We instantiate this framework for a human-like embodiment, defining the high-level action space as a small set of 2D waypoints annotated on the current observation frame, each specifying a near-term goal position for a leaf joint (pelvis, head, hands). Waypoints are low-dimensional, visually interpretable, and easy to specify manually or to search over. We show that the lifted world model substantially outperforms searching directly in low-level joint space ($3.8\times$ lower mean joint error to the goal pose), while remaining more compute-efficient and generalizing to environments unseen by the policy.

---

### SWAN: World-Aware Adaptive Multimodal Networks for Runtime Variations

Authors: Jason Wu, Shir-Kang Scott Jin, Yuyang Yuan, Maggie Wigness, Lance M. Kaplan, Hang Qiu, Mani Srivastava
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.26181v2

Multimodal deep neural networks deployed in realistic environments must contend with runtime variations: changes in modality quality, overall input complexity, and available platform resources. Current networks struggle with such fluctuations -- adaptive networks cannot adhere to a strict compute budget, controller-based networks neglect to consider input complexity, and statically provisioned networks fail at all the above. Consequently, they do not extract maximum utility from the expended computational resources. We present SWAN (Sample and World-Aware Multimodal Network), the first adaptive multimodal network that accomplishes all three goals. SWAN employs a quality-aware controller to assign resources among modalities according to a variable user-specified maximum budget. Within this budget, an adaptive gating module further optimizes efficiency by scaling layer utilization according to sample complexity. For further gains, SWAN also employs a token dropping module that masks semantically irrelevant multimodal features before performing detections. We evaluate SWAN in the domain of autonomous driving with complex multi-object 3D detection, reducing FLOPs by up to 49% with minimal degradation.

---

### Evergreen: Efficient Claim Verification for Semantic Aggregates

Authors: Alexander W. Lee, Benjamin Han, Shayak Sen, Sam Yeom, Ugur Cetintemel, Anupam Datta
Published: 2026-04-28
Categories: cs.DB, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2604.26180v2

With recent semantic query processing engines, semantic aggregation has become a primitive operator, enabling the reduction of a relation into a natural language aggregate using an LLM. However, the resulting semantic aggregate may contain claims that are not grounded in the underlying relation. Verifying such claims is challenging: they often involve quantifiers, groupings, and comparisons over relations that far exceed LLM context windows and require a costly combination of semantic and symbolic processing. We present Evergreen, a system that recasts claim verification as a semantic query processing task with tailored optimizations and provenance capture. Evergreen compiles each claim into a declarative semantic verification query that can execute on the same query engine used to produce the aggregate. To reduce cost, Evergreen avoids unnecessary LLM calls through verification-aware optimizations, including early stopping, relevance sorting, and estimation with confidence sequences, as well as general-purpose optimizations for semantic queries, such as operator fusion, similarity filtering, and prompt caching. Each verdict is accompanied by citations that identify a minimal set of tuples justifying the result, with semantics based on semiring provenance for first-order logic. On a benchmark of production-inspired workloads over restaurant review and customer support datasets, Evergreen's optimized configurations occupy the entire cost-quality Pareto frontier. With a strong LLM, Evergreen preserves verification quality at an F1 of 0.94 while reducing cost by 3.1x relative to unoptimized verification; with a substantially weaker LLM, it surpasses the strongest external baseline's F1 (0.87 vs. 0.83) at 7.0x lower cost.

---

### CacheRAG: A Semantic Caching System for Retrieval-Augmented Generation in Knowledge Graph Question Answering

Authors: Yushi Sun, Lei Chen
Published: 2026-04-28
Categories: cs.DB, cs.CL
arXiv: https://arxiv.org/abs/2604.26176v4

The integration of Large Language Models (LLMs) with Retrieval-Augmented Generation (RAG) has significantly advanced Knowledge Graph Question Answering (KGQA). However, existing LLM-driven KGQA systems act as stateless planners, generating retrieval plans in isolation without exploiting historical query patterns: analogous to a database system that optimizes every query from scratch without a plan cache. This fundamental design flaw leads to schema hallucinations and limited retrieval coverage. We propose CacheRAG, a systematic cache-augmented architecture for LLM-based KGQA that transforms stateless planners into continual learners. Unlike traditional database plan caching (which optimizes for frequency), CacheRAG introduces three novel design principles tailored for LLM contexts: (1) Schema-agnostic user interface: A two-stage semantic parsing framework via Intermediate Semantic Representation (ISR) enables non-expert users to interact purely in natural language, while a Backend Adapter grounds the LLM with local schema context to compile executable physical queries safely. (2) Diversity-optimized cache retrieval: A two-layer hierarchical index (Domain $\rightarrow$ Aspect) coupled with Maximal Marginal Relevance (MMR) maximizes structural variety in cached examples, effectively mitigating reasoning homogeneity. (3) Bounded heuristic expansion: Deterministic depth and breadth subgraph operators with strict complexity guarantees significantly enhance retrieval recall without risking unbounded API execution. Extensive experiments on multiple benchmarks demonstrate that CacheRAG significantly outperforms state-of-the-art baselines (e.g., +13.2% accuracy and +17.5% truthfulness on the CRAG dataset).

---

### Why Domain Matters: A Preliminary Study of Domain Effects in Underwater Object Detection

Authors: Melanie Wille, Dimity Miller, Tobias Fischer, Scarlett Raine
Published: 2026-04-28
Categories: cs.CV, cs.LG, cs.RO
arXiv: https://arxiv.org/abs/2604.26174v1

Domain shift, where deviations between training and deployment data distributions degrade model performance, is a key challenge in underwater environments. Existing benchmarks testing performance for underwater domain shift simulate variability through synthetic style transfer. This fails to capture intrinsic scene factors such as visibility, illumination, scene composition, or acquisition factors, limiting analysis of real-world effects. We propose a labeling framework that defines underwater domains using measurable image, scene, and acquisition characteristics. Unlike prior benchmarks, it captures physically meaningful factors, enabling semantically consistent image grouping and supporting domain-specific evaluation of detection performance including failure analysis. We validate this on public datasets, showing systematic variations across domain factors and revealing hidden failure modes.

---

### Entropy Centroids as Intrinsic Rewards for Test-Time Scaling

Authors: Wenshuo Zhao, Qi Zhu, Xingshan Zeng, Fei Mi, Lifeng Shang, Yi R., Fung
Published: 2026-04-28
Categories: cs.LG, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2604.26173v2

An effective way to scale up test-time compute of large language models is to sample multiple responses and then select the best one, as in Grok Heavy and Gemini Deep Think. Existing selection methods often rely on external reward models, which requires training a strong reward model and introduces additional computation overhead. As an alternative, previous approaches have explored intrinsic signals, such as confidence and entropy, but these signals are noisy with naive aggregation. In this work, we observe that high-entropy tokens tend to cluster into consecutive groups during inference, providing a more stable notion of model uncertainty than individual tokens. Together, these clusters reveal temporal patterns of model uncertainty throughout the inference process. Motivated by this observation, we propose to use the temporal structure of uncertainty as an intrinsic reward. To this end, we first formalize the basic unit of segment-level uncertainty as the High Entropy Phase (HEP), a variable-length segment that begins at a high-entropy token and ends when consecutive low-entropy tokens appear. We then define the Entropy Centroid, inspired by the concept of the center of mass in physics, as the weighted average position of all HEPs along the trajectory. Intuitively, a lower centroid indicates early exploration followed by confident generation, which we find often corresponds to higher response quality. Based on this insight, we propose the Lowest Centroid method, which selects the response with the lowest entropy centroid among multiple candidates. Experiments on mathematics, code generation, logical reasoning, and agentic tasks, across model scales ranging from 14B to 480B, show that Lowest Centroid consistently outperforms existing baselines and delivers stable gains as model size increases. Code is available at https://github.com/hkust-nlp/entropy-centroid.

---

### Co-Learning Port-Hamiltonian Systems and Optimal Energy-Shaping Control

Authors: Ankur Kamboj, Biswadip Dey, Vaibhav Srivastava
Published: 2026-04-28
Categories: eess.SY, cs.AI, cs.LG, math.OC, stat.ML
arXiv: https://arxiv.org/abs/2604.26172v2

We develop a physics-informed learning framework for energy-shaping control of port-Hamiltonian (pH) systems from trajectory data. The proposed approach co-learns a pH system model and an optimal energy-balancing passivity-based controller (EB-PBC) through alternating optimization with policy-aware data collection. At each iteration, the system model is refined using trajectory data collected under the current control policy, and the controller is re-optimized on the updated model. Both components are parameterized by neural networks that embed the pH dynamics and EB-PBC structure, ensuring interpretability in terms of energy interactions. The learned controller renders the closed-loop system inherently passive and provably stable, and exploits passive plant dynamics without canceling the natural potential. A dissipation regularization enforces strict energy decay during training, thereby enhancing robustness to sim-to-real gaps. The proposed framework is validated on state-regulation and swing-up tasks for planar and torsional pendulum systems.

---

### EvoSelect: Data-Efficient LLM Evolution for Targeted Task Adaptation

Authors: Ting-Wei Li, Sirui Chen, Jiaru Zou, Yingbing Huang, Tianxin Wei, Jingrui He, Hanghang Tong
Published: 2026-04-28
Categories: cs.CL
arXiv: https://arxiv.org/abs/2604.26170v2

Adapting large language models (LLMs) to a targeted task efficiently and effectively remains a fundamental challenge. Such adaptation often requires iteratively improving the model toward a targeted task, yet collecting high-quality human-labeled data to support this process is costly and difficult to scale. As a result, synthetic data generation has emerged as a flexible and scalable alternative. One straightforward approach is through an iterative generation-training loop, where candidate data are synthesized through an external generator, the model is updated using these data and the process is repeated over iterations. However, generated samples can be noisy, highly redundant, or even misaligned with the targeted task distribution. Training indiscriminately on such data can dilute useful learning signals and even degrade model performance. To address this, we introduce a refined paradigm, namely an iterative generation-selection-training loop, which incorporates a selection step prior to model updates. Building on this paradigm, we propose EvoSelect, a data-efficient framework to evolve LLM effectively. Given candidate samples produced by the data generator, EvoSelect selects training data by jointly modeling targeted task alignment and diversity. We estimate task relevance through optimal transport with proxy gradient representations, which quantifies how well candidate samples align with the targeted task distribution. To mitigate redundancy, we incorporate a diversification mechanism that promotes coverage of complementary training samples. By interleaving alignment and diversification, EvoSelect enables progressive LLM evolution toward targeted tasks. Extensive experiments on various benchmarks demonstrate that with either weak or strong data generators, EvoSelect consistently improves adaptation efficacy over existing data selection methods.

---

### EvoSelect: Data-Efficient LLM Evolution for Targeted Task Adaptation

Authors: Ting-Wei Li, Sirui Chen, Jiaru Zou, Yingbing Huang, Tianxin Wei, Jingrui He, Hanghang Tong
Published: 2026-04-28
Categories: cs.CL
arXiv: https://arxiv.org/abs/2604.26170v1

Adapting large language models (LLMs) to a targeted task efficiently and effectively remains a fundamental challenge. Such adaptation often requires iteratively improving the model toward a targeted task, yet collecting high-quality human-labeled data to support this process is costly and difficult to scale. As a result, synthetic data generation has emerged as a flexible and scalable alternative. One straightforward approach is through an iterative generation-training loop, where candidate data are synthesized through an external generator, the model is updated using these data and the process is repeated over iterations. However, generated samples can be noisy, highly redundant, or even misaligned with the targeted task distribution. Training indiscriminately on such data can dilute useful learning signals and even degrade model performance. To address this, we introduce a refined paradigm, namely an iterative generation-selection-training loop, which incorporates a selection step prior to model updates. Building on this paradigm, we propose EvoSelect, a data-efficient framework to evolve LLM effectively. Given candidate samples produced by the data generator, EvoSelect selects training data by jointly modeling targeted task alignment and diversity. We estimate task relevance through optimal transport with proxy gradient representations, which quantifies how well candidate samples align with the targeted task distribution. To mitigate redundancy, we incorporate a diversification mechanism that promotes coverage of complementary training samples. By interleaving alignment and diversification, EvoSelect enables progressive LLM evolution toward targeted tasks. Extensive experiments on various benchmarks demonstrate that with either weak or strong data generators, EvoSelect consistently improves adaptation efficacy over existing data selection methods.

---

### Budget-Constrained Causal Bandits: Bridging Uplift Modeling and Sequential Decision-Making

Authors: Abhirami Pillai
Published: 2026-04-28
Categories: cs.LG, econ.EM, stat.ML
arXiv: https://arxiv.org/abs/2604.26169v2

Treatment allocation under budget constraints is a central challenge in digital advertising. The standard approach trains an offline uplift model on historical data, then solves a constrained optimization to allocate budget. This fails in cold-start settings where little historical data exists. We propose Budget-Constrained Causal Bandits (BCCB), an online framework that learns which users respond to ads while simultaneously spending the budget. BCCB unifies three components: learning individual-level treatment effects, exploring users whose response is uncertain, and pacing the budget over time. We derive the per-arrival decision rule as the KKT condition of a Lagrangian relaxation of the budgeted causal-allocation objective, providing a principled foundation for the algorithm. We evaluate on the Criteo Uplift dataset using 20 random seeds with paired statistical tests. Our central finding is a data-efficiency crossover at n = 7,500 historical observations (paired one-sided t-test, p = 0.043): below this threshold, offline pipelines either fail or produce unreliable allocations, while BCCB operates from the first user. BCCB exhibits 2-4x lower run-to-run variance than offline methods and outperforms all four online baselines (Thompson Sampling, budgeted Thompson Sampling, HTE Greedy, and Uplifting Bandits) at every budget level tested (p < 0.001). These results give practitioners a concrete decision rule for choosing between offline and online paradigms.

---

### Budget-Constrained Causal Bandits: Bridging Uplift Modeling and Sequential Decision-Making

Authors: Abhirami Pillai
Published: 2026-04-28
Categories: cs.LG, econ.EM, stat.ML
arXiv: https://arxiv.org/abs/2604.26169v1

Treatment allocation under budget constraints is a central challenge in digital advertising: advertisers must decide which users to show ads to while spending a limited budget wisely. The standard approach follows a two-stage offline pipeline - first collect historical data to estimate heterogeneous treatment effects (HTE), then solve a constrained optimization to allocate the budget. This works well with abundant data, but fails in cold-start settings such as new campaigns, new markets, or new customer segments where little historical data exists. We propose Budget-Constrained Causal Bandits (BCCB), an online framework that learns which users respond to ads while simultaneously spending the budget, making treatment decisions one user at a time. BCCB unifies three components into a single sequential process: learning individual-level ad effectiveness, exploring users whose response is uncertain, and pacing the budget over time. We evaluated on the Criteo Uplift dataset, a large-scale advertising dataset from a real randomized controlled trial. Our key finding is a data-efficiency crossover: offline methods require approximately 10,000 historical observations to produce reliable results, while BCCB operates effectively from the very first user. Furthermore, BCCB exhibits 3-5x lower performance variance between runs, making it more practical for real campaign planning. Among purely online methods, BCCB consistently outperforms standard Thompson Sampling, budgeted Thompson Sampling, and greedy HTE estimation across all budget levels tested.

---

### Test-Time Safety Alignment

Authors: Baturay Saglam, Dionysis Kalogerias
Published: 2026-04-28
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2604.26167v1

Recent work has shown that a model's input word embeddings can serve as effective control variables for steering its behavior toward outputs that satisfy desired properties. However, this has only been demonstrated for pretrained text-completion models on the relatively simple objective of reducing surface-level profanity in short continuations. A natural and practically important question is how well input embeddings can control aligned models, which produce an imbalanced bimodal refuse-or-comply output distribution rather than the smooth distribution characteristic of open-ended generation. We explore this in the context of safety, showing that input word embeddings can be optimized in a sub-lexical manner to minimize the semantic harmfulness of aligned model responses. Our approach uses zeroth-order gradient estimation of a black-box text-moderation API with respect to the input embeddings, and then applies gradient descent on these embeddings to minimize the harmfulness of the generated text. Experiments show that the proposed method can neutralize every safety-flagged response on standard safety benchmarks.

---

### Fitting Large Nonlinear Mixed Effects Models Using Variational Expectation Maximization

Authors: Mohamed Tarek, Pedro Afonso
Published: 2026-04-28
Categories: stat.ME, cs.CE, cs.LG, cs.MS, stat.CO
arXiv: https://arxiv.org/abs/2604.26160v2

Nonlinear Mixed Effects (NLME) models are widely used in pharmacometrics and related fields to analyze hierarchical and longitudinal data. However, as the number of parameters and random effects increases, traditional methods for maximizing the marginal likelihood become computationally expensive. This paper explores the Variational Expectation Maximization (VEM) algorithm, a scalable alternative for fitting NLME models. Originally introduced in the context of probabilistic graphical models and later popularized through variational autoencoders, VEM has not been extensively applied to NLME modeling. By leveraging flexible variational families and reverse-mode automatic differentiation, VEM can efficiently maximize the marginal likelihood, scaling to NLME models with over 15,000 population parameters. This work provides a detailed description of VEM, compares it to other NLME fitting algorithms, and highlights its scalability through computational experiments. Using the Pumas statistical software, we fit two test models: 1) a standard warfarin model, and 2) an unnecessarily over-parameterized DeepNLME Friberg model with 15,410 population parameters and 16 random effects. The warfarin model was fitted to completion to demonstrate the correctness of VEM, while the DeepNLME Friberg model instead demonstrates VEM's scalability on a toy but large model. VEM improves the log likelihood steadily over hundreds of iterations at a practical per-iteration cost, while FOCE fails to complete even one iteration within a day. The model is deliberately over-parameterized for its small dataset and over-fits it, so what this experiment establishes is that VEM optimizes the objective of a model of this size at a practical cost. Applying VEM to large models that are genuinely useful is left to future work.

---

### Fitting Large Nonlinear Mixed Effects Models Using Variational Expectation Maximization

Authors: Mohamed Tarek, Pedro Afonso
Published: 2026-04-28
Categories: stat.ME, cs.CE, cs.LG, cs.MS, stat.CO
arXiv: https://arxiv.org/abs/2604.26160v1

Nonlinear Mixed Effects models (NLME) models are widely used in pharmacometrics and related fields to analyze hierarchical and longitudinal data. However, as the number of parameters and random effects increases, traditional methods for maximizing the marginal likelihood become computationally expensive. This paper explores the Variational Expectation Maximization (VEM) algorithm, a scalable alternative for fitting NLME models. Originally introduced in the context of probabilistic graphical models and later popularized through variational autoencoders, VEM has not been extensively applied to NLME modeling. By leveraging flexible variational families and reverse-mode automatic differentiation, VEM can efficiently maximize the marginal likelihood, scaling to NLME models with over 15,000 population parameters. This work provides a detailed description of VEM, compares it to other NLME fitting algorithms, and highlights its scalability through computational experiments. Using the Pumas statistical software, we fit two test models: 1) a standard warfarin model, and 2) a DeepNLME Friberg model with 15,410 population parameters and 16 random effects. The warfarin model was fitted to completion to demonstrate the correctness of VEM, while the DeepNLME Friberg model was fitted for a limited number of iterations to measure the time per iteration and demonstrate VEM's scalability.

---

### Structural Generalization on SLOG without Hand-Written Rules

Authors: Zichao Wei
Published: 2026-04-28
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2604.26157v4

Structural generalization in semantic parsing requires systems to apply learned compositional rules to novel structural combinations. Existing approaches either rely on hand-written algebraic rules (AM-Parser) or fail to generalize structurally (Transformer-based models). We present an alternative requiring no hand-written compositional rules, based on a neural cellular automaton (NCA) with a discrete bottleneck: all compositional rules are learned from data through local iteration. On the SLOG benchmark, the system achieves an overall accuracy of $67.3 \pm 0.2\%$ across 10 seeds (AM-Parser: $70.8 \pm 4.3\%$), with 11 of 17 structural generalization categories at $100\%$ type-exact match, including three where AM-Parser scores $0$--$74\%$. Analysis reveals that all 5,539 failure instances reduce to exactly two mechanisms: novel combinations of wh-extraction context with reduced verb types, and modifiers appearing on the subject side of verbs. When we decompose results by CCG structural features, each sub-pattern either succeeds on all instances or fails on all. Intermediate scores (e.g., $41.4\%$) are mixtures of structurally distinct CCG patterns, not partial generalization. These results suggest that CCG directed types provide higher resolution than SLOG's phenomenon-level categories for characterizing structural generalization, and that the success/failure boundary is determined by the coverage of directed operations in the training data.

---

### Structural Generalization on SLOG without Hand-Written Rules

Authors: Zichao Wei
Published: 2026-04-28
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2604.26157v3

Structural generalization in semantic parsing requires systems to apply learned compositional rules to novel structural combinations. Existing approaches either rely on hand-written algebraic rules (AM-Parser) or fail to generalize structurally (Transformer-based models). We present an alternative requiring no hand-written compositional rules, based on a neural cellular automaton (NCA) with a discrete bottleneck: all compositional rules are learned from data through local iteration. On the SLOG benchmark, the system achieves an overall accuracy of $67.3 \pm 0.2\%$ across 10 seeds (AM-Parser: $70.8 \pm 4.3\%$), with 11 of 17 structural generalization categories at $100\%$ type-exact match, including three where AM-Parser scores $0$--$74\%$. Analysis reveals that all 5,539 failure instances reduce to exactly two mechanisms: novel combinations of wh-extraction context with reduced verb types, and modifiers appearing on the subject side of verbs. When we decompose results by CCG structural features, each sub-pattern either succeeds on all instances or fails on all. Intermediate scores (e.g., $41.4\%$) are mixtures of structurally distinct CCG patterns, not partial generalization. These results suggest that CCG directed types provide higher resolution than SLOG's phenomenon-level categories for characterizing structural generalization, and that the success/failure boundary is determined by the coverage of directed operations in the training data.

---

### Beyond Screenshots: Evaluating VLMs' Understanding of UI Animations

Authors: Chen Liang, Xirui Jiang, Naihao Deng, Eytan Adar, Anhong Guo
Published: 2026-04-28
Categories: cs.HC, cs.CL
arXiv: https://arxiv.org/abs/2604.26148v3

AI agents operating on user interfaces must understand how interfaces communicate state and feedback to act reliably. As a core communicative modality, animations are increasingly used in modern interfaces, serving critical functional purposes beyond mere aesthetics. Thus, understanding UI animation is essential for comprehensive interface interpretation. However, recent studies of Vision Language Models (VLMs) for UI understanding have focused primarily on static screenshots, leaving it unclear how well these models handle dynamic UI animations. To address this gap, we created AniMINT, a novel dataset of 300 densely annotated UI animation videos. We systematically evaluate state-of-the-art VLMs on UI animation understanding, including their abilities to perceive the animation effects, identify animation purposes, and interpret animation meaning. Our results show that VLMs can reliably detect primitive motion. However, their high-level animation interpretation remains inconsistent, with substantial gaps relative to human performance. Finally, we use Motion, Context, and Perceptual Cues (MCPC) to probe factors affecting VLM performance, revealing key bottlenecks and directions for future improvement.

---

### Beyond Screenshots: Evaluating VLMs' Understanding of UI Animations

Authors: Chen Liang, Xirui Jiang, Naihao Deng, Eytan Adar, Anhong Guo
Published: 2026-04-28
Categories: cs.HC, cs.CL
arXiv: https://arxiv.org/abs/2604.26148v1

AI agents operating on user interfaces must understand how interfaces communicate state and feedback to act reliably. As a core communicative modality, animations are increasingly used in modern interfaces, serving critical functional purposes beyond mere aesthetics. Thus, understanding UI animation is essential for comprehensive interface interpretation. However, recent studies of Vision Language Models (VLMs) for UI understanding have focused primarily on static screenshots, leaving it unclear how well these models handle dynamic UI animations. To address this gap, we created AniMINT, a novel dataset of 300 densely annotated UI animation videos. We systematically evaluate state-of-the-art VLMs on UI animation understanding, including their abilities to perceive the animation effects, identify animation purposes, and interpret animation meaning. Our results show that VLMs can reliably detect primitive motion. However, their high-level animation interpretation remains inconsistent, with substantial gaps relative to human performance. Finally, we use Motion, Context, and Perceptual Cues (MCPC) to probe factors affecting VLM performance, revealing key bottlenecks and directions for future improvement.

---

### A Data-Centric Framework for Intraoperative Fluorescence Lifetime Imaging for Glioma Surgical Guidance

Authors: Silvia Noble Anbunesan, Mohamed Abul Hassan, Jinyi Qi, Lisanne Kraft, Han Sung Lee, Orin Bloch, Laura Marcu
Published: 2026-04-28
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2604.26147v1

Accurate intraoperative assessment of glioma infiltration is essential for maximizing tumor resection while preserving functional brain tissue. Fluorescence lifetime imaging (FLIm) offers real-time, label-free biochemical contrast, but its clinical utility is challenged by biological heterogeneity, class imbalance, and variability in histopathological labeling. We present a data-centric AI (DC-AI) framework that integrates confident learning (CL), class refinement, and targeted label evaluation to develop a robust multi-class FLIm classifier for glioblastoma (GBM) resection margins. FLIm data were collected from 192 tissue margins across 31 newly diagnosed IDH-wildtype GBM patients and initially labeled into seven tumor cellularity classes by an expert neuropathologist. CL was applied to quantify FLIm point-level confidence, identify label inconsistencies, and guide iterative class merging into a three-class scheme ("low", "moderate", "high"). The resulting high-fidelity dataset enabled training a model that achieved 96% accuracy in the three-class task. SHAP analysis revealed class-specific FLIm feature importance, highlighting distinct optical signatures across the infiltration spectrum. Targeted FLIm analysis further identified biological (e.g., gray matter composition) and acquisition-related (e.g., blood contamination) contributors to low-confidence predictions. Blinded re-evaluation of margins flagged by CL demonstrated intra-pathologist variability, underscoring the value of selective relabeling rather than exhaustive review. Together, these findings demonstrate that a DC-AI framework can systematically improve data reliability, enhance model robustness, and refine biological interpretation of FLIm signals, supporting the development of clinically actionable optical tools for real-time glioma margin assessment.

---

### Ceci n'est pas une explication: Evaluating Explanation Failures as Explainability Pitfalls in Language Learning Systems

Authors: Ben Knight, Wm. Matthew Kennedy, Danielle Carvalho, Isaac Pattis, James Edgell
Published: 2026-04-28
Categories: cs.HC, cs.AI
arXiv: https://arxiv.org/abs/2604.26145v2

AI-powered language learning tools increasingly provide instant, personalised feedback to millions of learners worldwide. However, this feedback can fail in ways that are difficult for learners--and even teachers--to detect, potentially reinforcing misconceptions and eroding learning outcomes over extended use. We present a portion of L2-Bench, a benchmark for evaluating AI systems in language education that includes (but is not limited to) six critical dimensions of effective feedback: diagnostic accuracy, awareness of appropriacy, causes of error, prioritisation, guidance for improvement, and supporting self-regulation. We analyse how AI systems can fail with respect to these dimensions. These failures, which we argue are conducive to "explainability pitfalls," are AI-generated explanations that appear helpful on the surface but are fundamentally flawed, increasing the risk of attainment, human-AI interaction, and socioaffective harms. We discuss how the specific context of language learning amplifies these risks and outline open questions we believe merit more attention when designing evaluation frameworks specifically. Our analysis aims to expand the community's understanding of both the typology of explainability pitfalls and the contextual dynamics in which they may occur in order to encourage AI developers to better design safe, trustworthy, and effective AI explanations.

---

### Mixture of Experts Framework in Machine Learning Interatomic Potentials for Atomistic Simulations

Authors: Gabriel de Miranda Nascimento, Marc L. Descoteaux, Laura Zichi, Chuin Wei Tan, William C. Witt, Nicola Molinari, Sriteja Mantha, Daniil Kitchaev, Mordechai Kornbluth, Karim Gadelrab, Charles Tuffile, Boris Kozinsky
Published: 2026-04-28
Categories: physics.comp-ph, cond-mat.mtrl-sci, cs.LG
arXiv: https://arxiv.org/abs/2604.26143v1

First-principles atomistic simulations are essential for understanding complex material phenomena but are fundamentally limited by their computational cost. While Machine Learning Interatomic Potentials (MLIPs) have drastically improved cost for a given accuracy, their inference cost remains a bottleneck for massive systems or long timescales. To address this, we introduce a multifidelity "Mixture-of-Experts" framework based on the E(3)-equivariant Allegro architecture. Our method spatially partitions the simulation domain into a chemically complex region (e.g., reactive interfaces) and a simple region (e.g., bulk lattice), assigning models of varying capacity to each. Among the challenges in such static domain decomposition, the mechanical mismatch between models at the interface is particularly critical, as it can generate artificial stress fields and instability. We address this challenge with a co-training strategy in which the loss function includes agreement constraints -- penalties on per-atom energy and force discrepancies between models evaluated on shared bulk environments -- forcing the independent models to learn a consistent physical description of the bulk material. We validate this approach on a realistic Pt+CO catalytic system, demonstrating that the co-trained models maintain exact energy conservation, align their bulk mechanical response (e.g., equation of state and bulk modulus), and achieve predictive accuracy comparable to a full high-fidelity simulation at more than twice the computational speed.

---

### ImproBR: Bug Report Improver Using LLMs

Authors: Emre Furkan Akyol, Mehmet Dedeler, Eray Tüzün
Published: 2026-04-28
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2604.26142v1

Bug tracking systems play a crucial role in software maintenance, yet developers frequently struggle with low-quality user-submitted reports that omit essential details such as Steps to Reproduce (S2R), Observed Behavior (OB), and Expected Behavior (EB). We propose ImproBR, an LLM-based pipeline that automatically detects and improves bug reports by addressing missing, incomplete, and ambiguous S2R, OB, and EB sections. ImproBR employs a hybrid detector combining fine-tuned DistilBERT, heuristic analysis, and an LLM analyzer, guided by GPT-4o mini with section-specific few-shot prompts and a Retrieval-Augmented Generation (RAG) pipeline grounded in Minecraft Wiki domain knowledge. Evaluated on Mojira, ImproBR improved structural completeness from 7.9% to 96.4%, more than doubled the proportion of executable S2R from 28.8% to 67.6%, and raised fully reproducible bug reports from 1 to 13 across 139 challenging real-world reports.

---

### HIVE: Hidden-Evidence Verification for Hallucination Detection in Diffusion Large Language Models

Authors: Guoshenghui Zhao, Tan Yu, Weijie Zhao
Published: 2026-04-28
Categories: cs.CL
arXiv: https://arxiv.org/abs/2604.26139v3

Diffusion large language models generate text through iterative denoising, exposing hidden trajectories that may contain reliability signals beyond the final output. We propose HIVE, which compresses trajectory hidden states, selects informative step-layer evidence, and conditions a verifier through continuous prefix embeddings to produce a hallucination score and structured diagnostics. Across two D-LLMs and three QA benchmarks, HIVE outperforms eight established baselines and a verifier-backbone-matched text-only control in all six settings. Relative to text-only verification, hidden-evidence conditioning improves AUROC by 1.73--4.60 points and AUPRC by 1.10--3.62 points, with average gains of 3.15 and 2.28 points, respectively. Ablations, evidence interventions, and cross-dataset transfer further support the complementary value of fine-grained hidden trajectory evidence.

---

### HIVE: Hidden-Evidence Verification for Hallucination Detection in Diffusion Large Language Models

Authors: Guoshenghui Zhao, Tan Yu, Weijie Zhao
Published: 2026-04-28
Categories: cs.CL
arXiv: https://arxiv.org/abs/2604.26139v2

Diffusion large language models generate text through multi-step denoising, where hallucination signals may emerge throughout the trajectory rather than only in the final output. Existing detectors mainly rely on output uncertainty or coarse trace statistics, which often fail to capture the richer hidden dynamics of D-LLMs. We propose HIVE, a hidden-evidence verification framework that extracts compressed hidden evidence from denoising trajectories, selects informative step-layer evidence, and conditions a verifier language model on the selected evidence through prefix embeddings. HIVE produces both a continuous hallucination score from verifier decision logits and structured verification outputs, including hallucination types, evidence pairs, and short rationales. Across two D-LLMs and three QA benchmarks, HIVE consistently outperforms eight strong baselines and achieves up to 0.9236 AUROC and 0.9537 AUPRC. Ablation studies further confirm the importance of hidden-evidence conditioning, learned evidence selection, two-stream evidence representation, and step-layer embeddings. These results suggest that selected hidden evidence from denoising trajectories provides a stronger and more usable hallucination signal than output-only uncertainty or coarse trace statistics.

---

### MixerCA: An Efficient and Accurate Model for High-Performance Hyperspectral Image Classification

Authors: Mohammed Q. Alkhatib, Ali Jamali
Published: 2026-04-28
Categories: cs.CV
arXiv: https://arxiv.org/abs/2604.26138v1

Over the past decade, hyperspectral image (HSI) classification has drawn considerable interest due to HSIs' ability to effectively distinguish terrestrial objects by capturing detailed, continuous spectral information. The strong performance of recent deep learning techniques in tasks like image classification and semantic segmentation has led to their growing use in HSI classification, due to their ability to capture complex spatial and spectral features more effectively than traditional methods. This paper presents MixerCA, a novel lightweight model for HSI classification that leverages depthwise convolution and a self-attention mechanism. MixerCA integrates depth-wise convolutions, token and channel mixing, and coordinate attention into a unified structure to decouple spatial and channel interactions, maintain consistent resolution throughout the network, and directly process HSI patches. Extensive experiments on four hyperspectral benchmark datasets reveal MixerCA's clear advantages over several competing algorithms, including 2D-CNN, 3D-CNN, Tri-CNN, HybridSN, ViT, and Swin Transformer. The source code is publicly available at https://github.com/mqalkhatib/MixerCA.

---

### One Voice, Many Tongues: Cross-Lingual Voice Cloning for Scientific Speech

Authors: Amanuel Gizachew Abebe, Yasmin Moslem
Published: 2026-04-28
Categories: eess.AS, cs.CL
arXiv: https://arxiv.org/abs/2604.26136v2

Preserving a speaker's voice identity while generating speech in a different language remains a fundamental challenge in spoken language technology, particularly in specialized domains such as scientific communication. In this paper, we address this challenge through our system submission to the International Conference on Spoken Language Translation (IWSLT 2026), the Cross-Lingual Voice Cloning shared task. First, we evaluate several state-of-the-art voice cloning models for cross-lingual speech generation of scientific texts in Arabic, Chinese, and French. Then, we build voice cloning systems based on the OmniVoice foundation model. We employ data augmentation via multi-model ensemble distillation from the ACL 60/60 corpus. We investigate the effect of using this synthetic data for fine-tuning, demonstrating improvements in intelligibility (WER & CER) and speaker similarity (SIM), with gains varying across languages.

---

### Spatially-constrained clustering of geospatial features for heat vulnerability assessment of favelas in Rio de Janeiro

Authors: Baptiste Clemence, Thomas Hallopeau, Vanderlei Pascoal De Matos, Laurent Demagistri, Joris Guerin
Published: 2026-04-28
Categories: cs.LG
arXiv: https://arxiv.org/abs/2604.26133v1

Informal settlements face disproportionate exposure to climate-related health hazards. However, existing methodologies lack systematic approaches to link diverse settlement characteristics with environmental health outcomes. We develop a data-driven framework to assess heat vulnerability in Rio de Janeiro's favelas by combining spatially-constrained clustering with land surface temperature (LST) analysis. Using remote sensing and geospatial features, we identify two distinct favela typologies: recent, well-connected settlements on flat terrain (Cluster 0) and historical, poorly-connected communities on vegetated slopes (Cluster 1). Analysis of 16 extreme heat events reveals systematic temperature differences of 2--3$^\circ$C between clusters, with flat-terrain favelas experiencing significantly higher heat exposure. Our findings demonstrate that settlement morphology critically influences heat vulnerability, providing a replicable framework for targeted urban planning and public health interventions in informal settlements globally.

---

### Sparse Graph Learning from Sparse Data via Fiedler Number Maximization

Authors: Bahar Oveisgharan, Gene Cheung, Andrew Eckford
Published: 2026-04-28
Categories: eess.SP, cs.LG
arXiv: https://arxiv.org/abs/2604.26132v1

We aim to learn a sparse and connected graph from sparse data, where the number of observations K can be substantially smaller than the signal dimension N for signals x in R^N, and the underlying distribution is unknown. In this severely ill-posed setting, we incorporate Fiedler number (the second eigenvalue of the graph Laplacian matrix that quantifies connectedness) as a robust regularization term in the sparse graph learning objective. We first develop a greedy algorithm that iteratively selects one edge globally for weakening/removal to reduce the objective, leveraging eigenvalue perturbation theorems that bound the adverse effect of an edge change to the Fiedler number. Next, we design a parallel variant, based on the Cheeger's inequality, that recursively partitions an input graph into two sub-graphs using an approximate Cheeger cut to distributedly find an optimal edge. Simulation experiments show that Fiedler number maximization robustifies sparse graph estimates, outperforming previous sparse graph learning algorithms.

---

### reward-lens: A Mechanistic Interpretability Library for Reward Models

Authors: Mohammed Suhail B Nadaf
Published: 2026-04-28
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2604.26130v1

Every RLHF-trained language model is shaped by a reward model, yet the mechanistic interpretability toolkit -- logit lens, direct logit attribution, activation patching, sparse autoencoders -- was built for generative LLMs whose primitives all project onto a vocabulary unembedding. Reward models replace that with a scalar regression head, breaking each tool. We present reward-lens, an open-source library that ports this toolkit to reward models, organised around one observation: the reward head's weight vector $w_r$ is the natural axis for every interpretability question. The library provides a Reward Lens, component attribution, three-mode activation patching, a reward-hacking probe suite, TopK SAE feature attribution, cross-model comparison, and five theory-grounded extensions (distortion index, divergence-aware patching, misalignment cascade detection, reward-term conflict analysis, concept-vector analysis). A ten-method adapter protocol covers Llama, Mistral, Gemma-2, and ArmoRM multi-objective heads, with a generic adapter for any HuggingFace sequence classification model. We validate on two production reward models across ~695 RewardBench pairs. The central empirical finding is negative: linear attribution does not predict causal patching effects (mean Spearman $ρ= -0.256$ on Skywork, $-0.027$ on ArmoRM). The framework treats this disagreement as a property to expose, not a bug -- motivating a design that keeps observational and causal views first-class and directly comparable.

---

### Robust Representation Learning through Explicit Environment Modeling

Authors: Yuli Slavutsky, David M. Blei
Published: 2026-04-28
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2604.26128v1

We consider learning from labeled data collected across multiple environments, where the data distribution may vary across these environments. This problem is commonly approached from a causal perspective, seeking invariant representations that retain causal factors while discarding spurious ones. However, this framework assumes that the environment has no direct effect on the target. In contrast, we consider settings in which this assumption fails, but still aim to learn representations that support robust prediction on average across previously unseen environments. To this end, we study representations learned by explicitly modeling variation across environments and then marginalizing that variation out. We analyze the resulting representations and characterize when they are preferable to those learned by causal invariant-representation methods. We propose a concrete method based on generalized random-intercept models, a class of predictors in which such marginalization is possible, and study their generalization properties. Empirically, we show that these models outperform invariant-learning methods across a range of challenging settings.

---

### Application of Deep Reinforcement Learning to Event-Triggered Control for Networked Artificial Pancreas Systems

Authors: Junya Ikemoto, Satoshi Maruyama, Kazumune Hashimoto
Published: 2026-04-28
Categories: eess.SY, stat.ML
arXiv: https://arxiv.org/abs/2604.26126v3

This paper proposes a deep reinforcement learning (DRL)-based event-triggered controller design for networked artificial pancreas (AP) systems. Although existing DRL-based AP controllers typically assume periodic control updates, networked control systems (NCSs) require a reduction in communication frequency to achieve energy-efficient operation, which is directly tied to control updates. However, jointly learning both insulin dosing and update timing significantly increases the complexity of the learning problem. To alleviate this complexity, we develop a practical DRL-based controller design that avoids explicitly learning update timing by introducing a rule-based criterion defined by changes in blood glucose. As a result, decision-making occurs at irregular intervals, and the problem is naturally formulated as a semi-Markov decision process (SMDP), for which we extend a standard DRL algorithm. Numerical experiments demonstrate that the proposed method improves communication efficiency while maintaining control performance.

---

### Hierarchical Multi-Persona Induction from User Behavioral Logs: Learning Evidence-Grounded and Truthful Personas

Authors: Nayoung Choi, Haeyu Jeong, Changbong Kim, Hongjun Lim, Jinho D. Choi
Published: 2026-04-28
Categories: cs.AI
arXiv: https://arxiv.org/abs/2604.26120v1

Behavioral logs provide rich signals for user modeling, but are noisy and interleaved across diverse intents. Recent work uses LLMs to generate interpretable natural-language personas from user logs, yet evaluation often emphasizes downstream utility, providing limited assurance of persona quality itself. We propose a hierarchical framework that aggregates user actions into intent memories and induces multiple evidence-grounded personas by clustering and labeling these memories. We formulate persona induction as an optimization problem over persona quality-captured by cluster cohesion, persona-evidence alignment, and persona truthfulness-and train the persona model using a groupwise extension of Direct Preference Optimization (DPO). Experiments on a large-scale service log and two public datasets show that our method induces more coherent, evidence-grounded, and trustworthy personas, while also improving future interaction prediction.

---

### Sample Selection Using Multi-Task Autoencoders in Federated Learning with Non-IID Data

Authors: Emre Ardıç, Yakup Genç
Published: 2026-04-28
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2604.26116v1

Federated learning is a machine learning paradigm in which multiple devices collaboratively train a model under the supervision of a central server while ensuring data privacy. However, its performance is often hindered by redundant, malicious, or abnormal samples, leading to model degradation and inefficiency. To overcome these issues, we propose novel sample selection methods for image classification, employing a multitask autoencoder to estimate sample contributions through loss and feature analysis. Our approach incorporates unsupervised outlier detection, using one-class support vector machine (OCSVM), isolation forest (IF), and adaptive loss threshold (AT) methods managed by a central server to filter noisy samples on clients. We also propose a multi-class deep support vector data description (SVDD) loss controlled by a central server to enhance feature-based sample selection. We validate our methods on CIFAR10 and MNIST datasets across varying numbers of clients, non-IID distributions, and noise levels up to 40%. The results show significant accuracy improvements with loss-based sample selection, achieving gains of up to 7.02% on CIFAR10 with OCSVM and 1.83% on MNIST with AT. Additionally, our federated SVDD loss further improves feature-based sample selection, yielding accuracy gains of up to 0.99% on CIFAR10 with OCSVM. These results show the effectiveness of our methods in improving model accuracy across various client counts and noise conditions.

---

### Evaluating Strategic Reasoning in Forecasting Agents

Authors: Tom Liptay, Dan Schwarz, Rafael Poyiadzi, Jack Wildman, Nikos I. Bosse
Published: 2026-04-28
Categories: cs.AI
arXiv: https://arxiv.org/abs/2604.26106v1

Forecasting benchmarks produce accuracy leaderboards but little insight into why some forecasters are more accurate than others. We introduce Bench to the Future 2 (BTF-2), 1,417 pastcasting questions with a frozen 15M-document research corpus in which agents reproducibly research and forecast offline, producing full reasoning traces. BTF-2 detects accuracy differences of 0.004 Brier score, and can distinguish differential agent strengths in research vs. judgment. We build a forecaster 0.011 Brier more accurate than any single frontier agent, and use it to evaluate agent strategic reasoning without hindsight bias. We find the better forecaster differs primarily in its pre-mortem analysis of its blind spots and consideration of black swans. Expert human forecasters found the dominant strategic reasoning failures of frontier agents are in assessing political and business leaders' incentives, judging their likelihood to follow through on stated plans, and modeling institutional processes.

---

### AMMA: A Multi-Chiplet Memory-Centric Architecture for Low-Latency 1M Context Attention Serving

Authors: Zhongkai Yu, Haotian Ye, Chenyang Zhou, Ohm Rishabh Venkatachalam, Zaifeng Pan, Zhengding Hu, Junsung Kim, Won Woo Ro, Po-An Tsai, Shuyi Pei, Yangwook Kang, Yufei Ding
Published: 2026-04-28
Categories: cs.AR, cs.AI, cs.DC, cs.LG
arXiv: https://arxiv.org/abs/2604.26103v2

All current LLM serving systems place the GPU at the center, from production-level attention-FFN disaggregation to NVIDIA's Rubin GPU-LPU heterogeneous platform. Even academic PIM/PNM proposals still treat the GPU as the central hub for cross-device communication. Yet the GPU's compute-rich architecture is fundamentally mismatched with the memory-bound nature of decode-phase attention, inflating serving latency while wasting power and die area on idle compute units. The problem is compounded as reasoning and agentic workloads push context lengths toward one million tokens, making attention latency the primary user-facing bottleneck. To address these inefficiencies, we present AMMA, a multi-chiplet, memory-centric architecture for low-latency long-context attention. AMMA replaces GPU compute dies with HBM-PNM cubes, roughly doubling the available memory bandwidth to better serve memory-bound attention workloads. To translate this bandwidth into proportional performance gains, we introduce (i) a logic-die microarchitecture that fully exploits per-cube internal bandwidth for decode attention under a minimal power and area budget, (ii) a two-level hybrid parallelism scheme, and (iii) a reordered collective flow that reduces intra-chip die-to-die communication overhead. We further conduct a design-space exploration over per-cube compute power and intra-chip D2D link bandwidth, providing actionable guidance for hardware designers. Evaluations show that AMMA achieves 15.5X lower attention latency and 6.9X lower energy consumption compared with the NVIDIA H100.

---

### SWE-Edit: Rethinking Code Editing for Efficient SWE-Agent

Authors: Yikai Zhang, Jiaxin Pei, Kenan Li, Qirui Jin, Maoquan Wang, Jin Pan, Yu Kang, Shengyu Fu, Elsie Nallipogu, Junjie Hu, Yufan Huang, Zijian Jin
Published: 2026-04-28
Categories: cs.SE, cs.CL
arXiv: https://arxiv.org/abs/2604.26102v2

Large language model agents have made strong progress on software engineering, yet current systems suffer from a context coupling problem: the standard code editing interface conflates code inspection, modification planning, and edit execution within a single context window, forcing agents to interleave exploratory viewing with strictly formatted edit generation. Irrelevant context accumulates and edit reliability degrades. We propose SWE-Edit, which decomposes the editing interface into two specialized subagents: a Viewer that extracts task-relevant code on demand, and an Editor that executes modifications from high-level natural language plans -- letting the main agent focus on reasoning while delegating context-intensive operations to clean context windows. On SWE-Bench Verified, this decomposition raises resolve rate by 2.1 pp and cuts inference cost by 17.9%, with consistent gains across multiple reasoning-model families (Kimi-K2, MiniMax-M2.1, GLM-4.7). We further show that effective edit-format selection can be trained into a small model rather than requiring frontier-scale capacity: GRPO training on Qwen3-8B with an adaptive find-replace/whole-file-rewrite policy improves edit success by 12.5 pp and brings an 8B open-source editor to parity with GPT-5-nano on downstream SWE-Bench resolve rate. To enable rapid editor iteration, we release PR-Edit, a lightweight evaluation whose scores correlate strongly with SWE-Bench resolve rate. We release our code at https://github.com/microsoft/SWE-Edit.

---

### Momentum-Conserving Graph Neural Networks for Deformable Objects

Authors: Jiahong Wang, Logan Numerow, Stelian Coros, Christian Theobalt, Vahid Babaei, Bernhard Thomaszewski
Published: 2026-04-28
Categories: cs.LG, cs.AI, cs.GR
arXiv: https://arxiv.org/abs/2604.26097v1

Graph neural networks (GNNs) have emerged as a versatile and efficient option for modeling the dynamic behavior of deformable materials. While GNNs generalize readily to arbitrary shapes, mesh topologies, and material parameters, existing architectures struggle to correctly predict the temporal evolution of key physical quantities such as linear and angular momentum. In this work, we propose MomentumGNN -- a novel architecture designed to accurately track momentum by construction. Unlike existing GNNs that output unconstrained nodal accelerations, our model predicts per-edge stretching and bending impulses which guarantee the preservation of linear and angular momentum. We train our network in an unsupervised fashion using a physics-based loss, and we show that our method outperforms baselines in a number of common scenarios where momentum plays a pivotal role.

---

### Distill-Belief: Closed-Loop Inverse Source Localization and Characterization in Physical Fields

Authors: Yiwei Shi, Zixing Song, Mengyue Yang, Cunjia Liu, Weiru Liu
Published: 2026-04-28
Categories: cs.AI
arXiv: https://arxiv.org/abs/2604.26095v1

{Closed-loop inverse source localization and characterization (ISLC) requires a mobile agent to select measurements that localize sources and infer latent field parameters under strict time constraints.} {The core challenge lies in the belief-space objective: valid uncertainty estimation requires expensive Bayesian inference, whereas using fast learned belief model leads to reward hacking, in which the policy exploits approximation errors rather than actually reducing uncertainty.} {We propose \textbf{Distill-Belief}, a teacher--student framework that decouples correctness from efficiency. A Bayes-correct particle-filter teacher maintains the posterior and supplies a dense information-gain signal, while a compact student distills the posterior into belief statistics for control and an uncertainty certificate for stopping. At deployment, only the student is used, yielding constant per-step cost.} {Experiments on seven field modalities and two stress tests show that Distill-Belief consistently reduces sensing cost and improves success, posterior contraction, and estimation accuracy over baselines, while mitigating reward hacking.}

---

### Operating-Layer Controls for Onchain Language-Model Agents Under Real Capital

Authors: T. J. Barton, Chris Constantakis, Patti Hauseman, Annie Mous, Alaska Hoffman, Brian Bergeron, Hunter Goodreau
Published: 2026-04-28
Categories: cs.AI, cs.CE, cs.MA
arXiv: https://arxiv.org/abs/2604.26091v1

We study reliability in autonomous language-model agents that translate user mandates into validated tool actions under real capital. The setting is DX Terminal Pro, a 21-day deployment in which 3,505 user-funded agents traded real ETH in a bounded onchain market. Users configured vaults through structured controls and natural-language strategies, but only agents could choose normal buy/sell trades. The system produced 7.5M agent invocations, roughly 300K onchain actions, about $20M in volume, more than 5,000 ETH deployed, roughly 70B inference tokens, and 99.9% settlement success for policy-valid submitted transactions. Long-running agents accumulated thousands of sequential decisions, including 6,000+ prompt-state-action cycles for continuously active agents, yielding a large-scale trace from user mandate to rendered prompt, reasoning, validation, portfolio state, and settlement. Reliability did not come from the base model alone; it emerged from the operating layer around the model: prompt compilation, typed controls, policy validation, execution guards, memory design, and trace-level observability. Pre-launch testing exposed failures that text-only benchmarks rarely measure, including fabricated trading rules, fee paralysis, numeric anchoring, cadence trading, and misread tokenomics. Targeted harness changes reduced fabricated sell rules from 57% to 3%, reduced fee-led observations from 32.5% to below 10%, and increased capital deployment from 42.9% to 78.0% in an affected test population. We show that capital-managing agents should be evaluated across the full path from user mandate to prompt, validated action, and settlement.

---

### FruitProM-V2: Robust Probabilistic Maturity Estimation and Detection of Fruits and Vegetables

Authors: Rahul Harsha Cheppally, Sidharth Rai, Sudan Baral, Benjamin Vail, Ajay Sharda
Published: 2026-04-28
Categories: cs.CV, cs.AI, cs.RO
arXiv: https://arxiv.org/abs/2604.26084v1

Accurate fruit maturity identification is essential for determining harvest timing, as incorrect assessment directly affects yield and post-harvest quality. Although ripening is a continuous biological process, vision-based maturity estimation is typically formulated as a multi-class classification task, which imposes sharp boundaries between visually similar stages. To examine this limitation, we perform an annotation reliability study with two independent annotators on a held-out tomato dataset and observe disagreement concentrated near adjacent maturity stages. Motivated by this observation, we model maturity as a latent continuous variable and predict it probabilistically using a distributional detection head, converting the distribution into class probabilities through the cumulative distribution function (CDF). The proposed formulation maintains comparable performance to a standard detector under clean labels while better representing uncertainty. Furthermore, when controlled label noise is introduced during training, the probabilistic model demonstrates improved robustness relative to the baseline, indicating that explicitly modeling maturity uncertainty leads to more reliable visual maturity estimation.

---