# Research Papers — 2026-06 · bundle page 42 of 260

> Full-text markdown of papers #2051–2100 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-06/page/42
[← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/41.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-06/page/43.md)

### Supersede: Diagnosing and Training the Memory-Update Gap in LLM Agents

Authors: Vedant Patel
Published: 2026-06-25
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2606.27472v1

Large language model (LLM) agents operate over long, multi-session interactions in which facts change: a user moves, a price updates, a plan is revised. Acting correctly requires using the current value of a fact and discarding values that have been superseded. We isolate this ability on real conversational data and show that it is a distinct, unsolved failure. On the knowledge-update subset of LongMemEval, replacing an agent's full context with a bounded, self-maintained memory drops accuracy from 92% to 77% even on a frontier model (gpt-5.4), a gap that is statistically significant (paired McNemar p<0.005) and persists across model scale while full-context accuracy saturates near 92%. The bottleneck is therefore memory maintenance, not comprehension, and is not closed by a stronger model. We then ask whether this is merely an undersized memory, and find it is not: as the conversation grows 24x, accuracy falls further (from 68% to 28%), and granting the agent proportionally more memory yields no detectable recovery (28% to 28%, n=25). The failure scales with the length of the conversation, not the compression ratio. We release Supersede, an open reinforcement-learning environment (on the verifiers / prime-rl stack) that turns this measurement into a training signal: agents are rewarded for answering from the current value and penalized for stale ones. Finally, we close the loop and show the gap is trainable: GRPO fine-tuning a small open model (Qwen2.5-3B) on this environment nearly doubles its held-out supersession accuracy on real, unseen conversations (9.0% to 16.7%, a single run), along a monotonic checkpoint curve indicating the learned policy, not the harness, carries the gain. To our knowledge this is the first trainable environment whose reward targets temporal fact-currency, and the first evidence the supersession gap can be trained down, not only measured.

---

### Multi-Objective Molecular Generation with Frequency-Controlled Evolutionary Dynamics

Authors: Elia Colleoni, Paolo Guida, Didier Barradas-Bautista, William Lafayette Roberts
Published: 2026-06-25
Categories: cs.NE
arXiv: https://arxiv.org/abs/2606.27467v1

Molecule generation methods that leverage generative models have been successfully applied to drug discovery. However, they often require extensive pre-training, suffer statistical biases in the training data, and might suffer from limited interpretability of generated chemical structures. In this work, we introduce SpectralMol, an algorithm based on evolutionary computation that processes chemical structures as a compact matrix of Fourier coefficients, projected onto a fixed basis to generate position-wise latent vectors for SELFIES decoding. The NSGA-II algorithm enforces diversity and enable separate objective functions rather than collapsed objectives into a scalar reward. The quality of the algorithm was tested against standardized benchmarks. The results show comparable aggregate benchmark performance with a task-dependent profile: SpectralMol is strongest on several multi-parameter optimization tasks. The same benchmark was used to perform an ablation study to demonstrate the advantages of a structured latent matrix. Finally, method was tested on a realistic ClpP-targeted drug-discovery benchmark, comparing it with the reinforcement-learning-based model under a fixed oracle-call budget. SpectralMol generates more docking hits and more diverse scaffolds while maintaining competitive physicochemical properties. The representation adopted in this work can cleanly separates scaffold-level modifications from localized substructure variations, as the former occur with perturbations of low-frequency Fourier modes and the latter with perturbations of high-frequency Fourier modes. The results support the evidence that frequency-controlled evolutionary dynamics provide an interpretable, efficient, and training-free route to multi-objective molecular design.

---

### The Decision Geometry of Covariance Estimation for the Global Minimum-Variance Portfolio under Heavy Tails

Authors: Xavier Fonseca
Published: 2026-06-25
Categories: stat.ML, cs.LG, q-fin.PM
arXiv: https://arxiv.org/abs/2606.27462v1

The global minimum-variance portfolio (GMVP) is the canonical decision built from an estimated covariance matrix, yet covariance estimators are universally evaluated by matrix-norm loss, which is not the object the decision depends on. We characterise exactly how covariance-estimation error maps into GMVP suboptimality. We prove an exact regret identity and a non-asymptotic bound showing decision regret depends on the estimation error only through its action on the portfolio weights, scaled by portfolio concentration and the conditioning of the true covariance. From this we derive the decision geometry: GMVP regret is invariant to a (p-1)-dimensional projection of the p^2-dimensional error matrix, with invariance to the covariance-scale direction as an exact special case. We then apply the framework to heavy-tailed returns (tail index kappa in (2,4)), establishing the regret convergence rate implied by the centred operator-norm rate, and confirm the theory on a skew-t/t-copula simulation design with pre-registered analysis. The decision-focused advantage is a sharper constant and a concentration discount rather than a faster rate; we report an honest high-conditioning boundary of the rate prediction. The results complement recent decision-focused learning approaches by supplying the exact estimation geometry and consistency theory they lack.

---

### Developmental approach reveals the statistical learning of Neural Language Models: Transformers generalize from the most abstract statistical patterns

Authors: Wang Bojun, Holly Jenkins, Elizabeth Wonnacott
Published: 2026-06-25
Categories: cs.CL
arXiv: https://arxiv.org/abs/2606.27460v1

In this study, we use a developmental approach to investigate the statistical learning and mental representation of neural language models (NLM). A series of Generative Transformer models are trained on a synthetic grammar. The model states are saved at multiple stages in the course of training. Through analyzing how the internal representations of these models change in the developmental path, we found that NLMs acquire the most abstract global statistical knowledge at the beginning of learning and later acquire the relatively local statistical dependencies. This learning path contains many over-generalizations from the very beginning and these over-generalizations are gradually constrained in the later stage of learning. Based on this observation, we propose a new framework to explain the statistical learning and language cognition of NLMs.

---

### Geometry-Conditioned Fourier Neural Operators for Cubic Nonlinear Schrodinger Dynamics on Periodic Domains

Authors: Emmanuel E. Oguadimma, Victory C. Obieke, Xueying Yu
Published: 2026-06-25
Categories: cs.LG, math.AP, math.NA
arXiv: https://arxiv.org/abs/2606.27459v2

We consider the cubic nonlinear Schrödinger (NLS) equation on two-dimensional flat tori with varying aspect ratios. In this formulation, the choice of aspect ratio governs the Fourier resonance structure, so rational and irrational geometries can exhibit different high-frequency cascade behaviors. We present a geometry-conditioned Fourier neural operator (FNO) for the cubic defocusing NLS equation, where the input consists of the real and imaginary parts of the solution together with the aspect-ratio parameter \(ω^2\). The model is trained to approximate the one-step solution operator and is evaluated on unseen trajectories generated from random-phase initial data using Fourier pseudospectral method. Our numerical experiments show that the learned operator captures the main solution dynamics on both tori and reproduces the distinct Sobolev norm behavior of the two geometries, with stronger \(H^2\)-growth on the rational torus and more constrained behavior on the irrational torus, consistent with the findings of \cite{hrabski2021energy}. We perform ablation studies to examine the roles of retained Fourier modes, activation functions, Fourier-layer depth, and explicit geometry conditioning. The results indicate that including $ω^2$ improves long-time predictive accuracy, especially for the rational geometry, and supports the use of geometry-aware neural operators for learning spectral-transfer phenomena in nonlinear dispersive partial differential equations.

---

### Cluster, Route, Escalate: Cascaded Framework for Cost-Aware LLM Serving

Authors: Yasmin Moslem, Magdalena Kacmajor, Vasudevan Nedumpozhimana, Ammar Abbas, Solmaz Panahi, David Lynch, Zhuangzhuang Nie, Alexandros Agapitos, Aleksandar Milenovic, Hongmeng Song, Yucheng Shi, Yue Pan et al.
Published: 2026-06-25
Categories: cs.PF, cs.CL
arXiv: https://arxiv.org/abs/2606.27457v1

Efficient deployment of large language models (LLMs) in production forces a trade-off between accuracy and cost. Operators often default to a single model that is either expensive for easy queries or insufficient for hard ones. To address this challenge, we propose a two-stage cascaded solution. Stage 1 clusters incoming queries and assigns each cluster to its most cost-effective model. The cost budget for this routing process is set by an interpretable hyperparameter, tuned offline. Stage 2 adds a quality estimation (QE) cascade; when an output from Stage 1 is judged low-quality, the query is escalated to a stronger model. This ensures only hard or low-confidence cases reach the expensive models. On the test datasets, the cascaded system retains 97-99% of the strongest model's accuracy while reducing Time Per Output Token (TPOT). It requires only task-correctness labels and adapts to changes in the model pool without manual reconfiguration.

---

### Directed Graph Topology Inference via Graph Filter Identification

Authors: Rasoul Shafipour, Andrei Buciulea, Santiago Segarra, Antonio G. Marques, Gonzalo Mateos
Published: 2026-06-25
Categories: stat.ML, cs.LG, cs.SI, eess.SP
arXiv: https://arxiv.org/abs/2606.27455v1

We address the problem of inferring a directed network from nodal measurements generated by linear diffusion dynamics on the sought graph. Observations are modeled as the outputs of a graph convolutional filter, i.e., a polynomial (with unknown coefficients) of a local diffusion graph-shift operator encoding the latent graph topology, excited with an ensemble of independent graph signals with arbitrarily-correlated nodal components. Unlike prior efforts that considered undirected graphs and white signal excitations, here the graph-shift operator and the observations' covariance matrix are not simultaneously diagonalizable. In this challenging context, we first rely on measurements of the output signals along with prior statistical information on the inputs to identify the diffusion filter. Such system identification problem involves solving a system of quadratic matrix equations, which we show is identifiable under spectral-diversity assumptions on the input covariances. For algorithmic purposes we recast it as a smooth quadratic minimization subject to Stiefel manifold constraints. Subsequent identification of the network topology given the graph filter estimate boils down to finding a sparse and structurally admissible shift that commutes with the given filter, thus, forcing the latter to be a polynomial in the sought graph-shift operator. A joint graph filter and topology identification algorithm is also proposed, which alternates between the aforementioned steps in a mutually reinforcing fashion to offer improved sample complexity. Numerical tests corroborate the effectiveness of the proposed algorithms in recovering synthetic digraphs and real-data case studies, and illustrate their potential utility on urban mobility analyses as well as portfolio optimization.

---

### Prism Transformer: Progressive Head Schedules for Hierarchical Attention Processing

Authors: Shubham Aggarwal
Published: 2026-06-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.27449v1

Multi-head attention conventionally partitions the hidden dimension equally across all heads at every layer, enforcing an identical representational subspace dimension (dh = dmodel/h) throughout the models depth. In this work, we identify this uniform allocation as a fundamental structural bottleneck: due to their restricted dimensional space, early-layer heads are unable to faithfully capture complex, high-dimensional contextual patterns. To resolve this, we introduce the Prism Transformer, a novel architectural paradigm that replaces the static, uniform head configuration with a progressive head schedule. By monotonically increasing the head count across layers, the Prism Transformer naturally establishes a local-to-global representational hierarchy: early layers leverage fewer, exceptionally wide heads to capture complex, local compositional patterns, while deep layers deploy many, narrow heads to decompose these patterns into specialized linguistic features. Crucially, this structural shift is parameter-neutral, compute-neutral, and introduces zero training or inference overhead, preserving identical weight matrices and FLOP budgets as the standard Transformer. Across three model scales (124M, 354M, and 757M), the Prism Transformer consistently outperforms uniform baselines, achieving consistent reductions in validation loss alongside consistent gains on downstream zero-shot benchmarks (including PIQA, HellaSwag, ARC-Easy, and WinoGrande). Our findings demonstrate that non-uniform subspace allocation unlocks latent capacity within the standard Transformer budget, enabling more effective use of model capacity.

---

### Learning in Markovian bandits with non-observable states and constrained decision epochs

Authors: Thomas Hira, Victor Boone, Urtzi Ayesta, Ina Maria Verloop
Published: 2026-06-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.27448v1

This paper studies the problem of regret minimization in Markovian bandits with \emph{non-observable states} and possibly \emph{constrained} decision epochs. The focus is restricted to a ``pure'' regret benchmark, that compares the performance of the learning algorithm to the best \emph{pure policy} which -- akin to optimal policies of stochastic bandits -- picks the optimal arm from start to finish without ever switching. We introduce a generalization of rested Markovian bandits, \emph{self-degrading Markovian bandits}, for which pure policies are always asymptotically optimal.We show that without prior knowledge on the underlying bandit, the regret of algorithms that switch arms rarely necessarily scales super-logarithmically for every bandit, i.e., as $ω(\log(T))$, where $T$ is the learning horizon. Despite the unreachability of the logarithmic regime, we design UCB-NOM, an optimistic algorithm inspired by UCB, of which the regret is nearly logarithmic. Lastly, we show that given prior knowledge on the Markovian bandit in the form of a bound on the bias functions of its arm, a proper instantiation of UCB-NOM achieves $O(\log(T))$ regret. We further show that this prior knowledge allows for a $O(\sqrt{T \log(T)})$ worst-case regret bound for UCB-NOM. Notably, our regret bounds do not depend on the number of states of the underlying Markov chains. Our findings suggest that the non-observability of states is a mild inconvenience in self-degrading Markovian bandits.

---

### Causal Connections: Leveraging Multilingual Fine-Tuning for Financial QA@FinCausal 2026

Authors: Akash Kumar Gautam, Serhii Hamotskyi, Christian Hänig
Published: 2026-06-25
Categories: cs.CL
arXiv: https://arxiv.org/abs/2606.27446v1

This paper describes team HSA_CORAL's submission to the FinCausal 2026 shared task on extracting cause-effect relations from financial narratives via extractive question answering in English and Spanish. We compare three modeling families: (i) encoder-only token tagging with multilingual BERT, (ii) encoder-decoder generation with multilingual BART, and (iii) decoder-only LLMs (Llama 3.1 and GPT variants) using prompt refinement, few-shot demonstrations, and supervised fine-tuning. Across settings, prompting and few-shot examples yield competitive performance, while supervised fine-tuning provides the largest gains. Our best system, GPT-4.1 Mini fine-tuned on combined English and Spanish training data, achieves a tied highest score on the English subtask (score 4.8140) and ranks third on Spanish (score 4.7753) under the shared task's LLM-as-a-judge metric. Overall, the results highlight the value of task-specific adaptation and multilingual fine-tuning for cross-lingual transfer in financial causality QA.

---

### SemCityLoc: Aerial 6DoF Localization Using Semantic 3D City Models

Authors: Jingfeng Mao, Xuyang Chen, Qilin Zhang, Oussema Dhaouadi, Guangming Wang, Brian Sheil, Daniel Cremers, Yan Xia, Olaf Wysocki
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27444v1

Aerial 6DoF localization typically relies on precise GNSS signals or radiometrically rich 3D reconstructions, limiting scalability and on-board deployment. We propose SemCityLoc, a semantic-geometric alignment system that reframes aerial pose estimation as structured surface registration between foundation-model-derived visual priors and standardized LoD-compliant 3D city models. Instead of matching sparse contours or dense texture, our method aligns semantic surfaces and monocular depth with lightweight semantic 3D building models, increasing pose discriminability in repetitive and occluded urban environments. To enable accurate evaluation, we introduce SemCityLockeD, the first real-world benchmark combining centimeter-accurate UAV poses with standardized LoD1--LoD3 semantic city models and challenging low-altitude imagery. Experiments demonstrate substantial improvements over existing map-based approaches, improving recall by up to 36% and reducing mean positional error from 9.89m to 2.62m in challenging urban canyons. Our results indicate that semantically structured geometry provides sufficient and scalable constraints for high-precision aerial localization without radiometric scene reconstructions. The code and data are available at https://albertchen98.github.io/SemCityLoc.

---

### When Does Personality Composition Matter for Multi-Agent LLM Teams?

Authors: Aryan Keluskar, Amrita Bhattacharjee, Huan Liu
Published: 2026-06-25
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2606.27443v2

Personality prompting shapes how large language models communicate, yet whether these behavioral shifts affect objective task outcomes remains under-explored. Prior work shows that agents prompted with low agreeableness produce adversarial language, while those prompted with high agreeableness become cooperative, but the relationship between communication style and task performance has not been systematically examined across multiple domains. In this work, we investigate whether personality composition matters for multi-agent team performance by manipulating personality traits across frontier LLMs on three task domains: structured coding, open-ended research collaboration, and competitive bargaining. We find that personality effects depend critically on task structure. In coding tasks, low agreeableness leads to large communication shifts that have little effect on milestone completion. In open-ended collaboration and bargaining, the same manipulation substantially degrades performance. We discuss implications for multi-agent system design and the limits of personality manipulation.

---

### PairSAE: Mechanistic Interpretability from Pair Representations in Protein Co-Folding

Authors: Giosue Migliorini, Aristofanis Rontogiannis, Grigori Guitchounts, Nicholas Franklin, Axel Elaldi, Olivia Viessmann
Published: 2026-06-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.27440v1

Foundation models for structural biology have achieved remarkable performance in predicting biomolecular structure and show promise for the design of proteins and small molecules. Yet understanding which internal features drive their outputs remains challenging. Standard sparse autoencoders (SAEs), effective on transformer-style sequence embeddings, do not transfer cleanly to pairformer-like architectures: naively operating on pairwise representations yields a quadratic blow-up of features and obscures concepts distributed jointly across sequence and pair representations. We introduce PairSAE, which summarizes pairwise tensors via an N-mode SVD into token-wise interaction roles, then uses a sparse autoencoder to learn a shared set of token-level features that decode into both sequence and pair representations. Evaluated on Boltz-2 activations for PLINDER protein-ligand complexes, PairSAE yields interpretable features that align with UniProt annotations and predict Boltz-2 affinity values. These results indicate that PairSAE links the latent space of foundation models for structural biology to interpretable structural concepts, clarifying what the model "knows" while avoiding pairformer-induced pitfalls that limit conventional SAEs.

---

### Unified Zero-Shot Time Series Forecasting: A Darts Foundation

Authors: Zhihao Dai, Dennis Bader, Alain Gysi
Published: 2026-06-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.27438v1

Since its initial release in 2020, Darts has become a widely used open-source Python library for time series analysis. A series of foundation models have recently claimed accuracy improvements in zero-shot forecasting, promising a paradigm shift from training custom models to harnessing pre-trained general-purpose forecasters. Foundation models, however, are often released as isolated packages with fragmented interfaces and limited interoperability with common tooling, making joint evaluation and integration within complete pipelines difficult. In Darts, we developed a unified $\texttt{FoundationModel}$ class collection (Chronos-2, TimesFM 2.5, TiRex, PatchTST-FM) that provides standardized, full-cycle forecasting interfaces with minimal external dependencies for integrating foundation models into the ecosystem. Existing Darts pipelines can now use foundation models with only a name change; new pipelines can use them for zero-shot or fine-tuned forecasting, uncertainty estimation, and backtesting, combined with data processing and evaluation tooling, all within a unified framework.

---

### GRAFT: Biological Graph and Hypergraph Benchmarks for Linked Gene Expression and Phenotypic Trait Prediction in Arabidopsis thaliana

Authors: Manuel Serna-Aguilera, Vanshika Jindal, Fiona L. Goggin, Jiamei Li, Aranyak Goswami, Alexander Bucksch, Suxing Liu, Khoa Luu
Published: 2026-06-25
Categories: q-bio.GN, cs.AI
arXiv: https://arxiv.org/abs/2606.27413v1

Understanding which genes control which traits in an organism remains one of the central challenges in biology. Despite significant advances in data collection technology, our ability to map genes to traits is still limited. This genome-to-phenome (G2P) challenge spans several problem domains, including plant breeding, and requires methods capable of reasoning over high-dimensional, heterogeneous, and biologically structured data. Current datasets and data repositories, however, are not well-equipped for this task. Current studies do not link gene expression and trait data, and most focus on very specific traits, limiting the breadth of possible correlations. To address this gap, we present the novel Gene-Graph Regression for Arabidopsis Functional Traits (GRAFT) dataset, a curated multi-modal dataset linking gene expression profiles with phenotypic trait measurements in Arabidopsis thaliana, a model organism in plant biology. GRAFT supports tasks such as phenotype prediction and interpretable graph learning. In addition, we benchmark conventional regression and explanatory baselines, including a biologically-informed hypergraph baseline, to validate gene-trait associations. To the best of our knowledge, this is the first dataset to provide multimodal gene information and heterogeneous trait or phenotype data for the same Arabidopsis thaliana specimens. With GRAFT, we aim to foster research to accurately understand the relationship between genotypes and phenotypes using gene information, higher-order gene pairings, and trait data from multiple sources.

---

### From Scene-Centric to Observer-Centric: Modeling Observer-Aware Relations for 3D Scene Graph Generation

Authors: Jingjun Sun, Chaowei Wang, Zhirui Liu, Jiaxu Tian, Ming Yang, Yaoxing Wang, Yan Di, Shan Gao
Published: 2026-06-25
Categories: cs.CV, cs.AI, cs.GR, eess.IV
arXiv: https://arxiv.org/abs/2606.27412v2

3D Scene Graph Generation (3DSGG) represents 3D scenes as structured object--relation--object graphs for spatial understanding. In observer-centric spatial perception, the same scene may be expressed under different local observer frames while its structure remains unchanged. However, existing models typically assume a fixed scene-aligned reference frame and may produce semantically inconsistent predictions when the scene is re-expressed in another observer frame. We attribute this failure to the heterogeneous frame dependency of relational predicates. Directional predicates such as $\textit{left}$, $\textit{front}$, $\textit{right}$, and $\textit{behind}$ are $\textbf{Observer-Dependent Relations}$, whereas most contact, support, and semantic predicates, such as $\textit{standing on}$ and $\textit{attached to}$, are approximately $\textbf{Observer-Independent Relations}$. Conventional models do not distinguish these frame responses, leading to degraded relation prediction under observer-frame reorientation. We introduce $\textbf{Observer-Aware Relations (OAR)}$, which combines observer-aware geometric encoding and relation specialization, supported by frame-stable object encoding, for unified multi-label predicate prediction. Experiments on 3DSSG show that OAR consistently outperforms baselines across controlled observer-frame reorientations without training-time frame-reorientation augmentation, while remaining competitive on the standard benchmark. The project page is available at https://oar-predicate.github.io/.

---

### Compression-Driven Anomaly Detection in Brain MRI Using an Interpretable Quantum Autoencoder

Authors: Santanu Ganguly, Xing Liang, Dimitrios Makris
Published: 2026-06-25
Categories: quant-ph, cs.AI, eess.IV
arXiv: https://arxiv.org/abs/2606.27411v1

We study a quantum autoencoder (QAE) for compression-driven anomaly detection in brain MRI data. The approach leverages angle encoding to map image patches into quantum states, followed by a variational encoder-decoder architecture trained to discard information via auxiliary trash qubits. Anomaly scores reflect the degree to which inputs resist compression relative to normal data, with higher scores corresponding to deviations from the learned normal manifold. Evaluated on publicly available brain MRI DICOM datasets, the method achieves a slice-level ROC-AUC of approximately 0.95 and a patch-level ROC-AUC of approximately 0.813, outperforming classical autoencoder and PCA baselines. Analysis of the learned parameters reveals a pronounced encoder-decoder asymmetry, where effective anomaly detection arises from structured information compression within the encoder rather than increased parameter magnitude or decoder expressivity. This results in a controlled compression-reconstruction trade-off with a clear operating regime that supports principled threshold selection. Qualitative evaluation further shows that the QAE produces spatially localized anomaly heatmaps aligned with tumorous regions. The results, supported by promising baseline performances, demonstrate that quantum autoencoders provide an interpretable and controllable mechanism for anomaly detection based on incompressibility with respect to a learned latent representation. This work highlights the potential of quantum autoencoders as a principled tool for studying compression dynamics in quantum machine learning, with promising implications for decision support in medical imaging workflows.

---

### DFM: Difference Feature Modeling with Text-Guided Gated Contrastive Loss for Remote Sensing Image Change Captioning

Authors: Yelin Wang, Zijia Song, Chuanguang Yang, Miaoyu Wang, Zhulin An, Libo Huang, Yongjun Xu
Published: 2026-06-25
Categories: eess.IV, cs.LG
arXiv: https://arxiv.org/abs/2606.27410v1

The primary goal of Remote Sensing Image Change Captioning (RSICC) is to automatically generate descriptions of changes between remote sensing images captured at different time points. Existing models still rely on a single autoregressive generation paradigm, which tends to prioritize learning easily generated vocabulary over capturing discriminative differences between images. To address this, we reframe the training paradigm and propose a novel Difference Feature Modeling (DFM) framework. Specifically, we introduce a Text-guided Gated Contrastive Loss (TGCL) to guide the vision encoder to extract critical features from a text-modal perspective. Additionally, we incorporate a pre-trained Change Detection model to transfer stable change detection knowledge. In order to further enhance the representation, we design a Joint Feature Modeling (JFM) module to achieve the fusion of multi-scale difference representations, thereby capturing comprehensive spatiotemporal variations between multi-temporal images. Extensive experiments on multiple datasets demonstrate the effectiveness of our approach.

---

### Delayed Verification Destabilizes Multi-Agent LLM Belief: Instability Thresholds and Optimal Corrector Placement

Authors: Igor Itkin
Published: 2026-06-25
Categories: cs.MA, cs.CL, cs.LG, eess.SY
arXiv: https://arxiv.org/abs/2606.27409v2

Multi-agent large language model (LLM) systems often rely on verifier and critic agents to suppress hallucinations, but verification is delayed. During this delay, false claims can propagate through the agent network. We model this process as delayed consensus on a graph with grounded corrector nodes. Spectral decomposition by the grounded Laplacian yields a closed-form stability threshold for the verification dose: correction that is too strong or too delayed can turn consensus into oscillation. The most unstable regime occurs when the communication and verification delays coincide; for delay two, the threshold is the inverse golden ratio. The same framework gives a supermodular placement objective and a greedy (1-1/e)-approximation rule for assigning a limited corrector budget to influential nodes. Experiments across five open models confirm the predicted dose-delay oscillations. By contrast, grounded factual answering makes truth an absorbing boundary and eliminates the effect, suggesting that the instability is specific to signed-belief tasks while grounded verification remains stabilizing

---

### Delayed Verification Destabilizes Multi-Agent LLM Belief: Instability Thresholds and Optimal Corrector Placement

Authors: Igor Itkin
Published: 2026-06-25
Categories: cs.MA, cs.CL, cs.LG, eess.SY
arXiv: https://arxiv.org/abs/2606.27409v1

Multi-agent large language model (LLM) systems often rely on verifier and critic agents to suppress hallucinations, but verification is delayed. During this delay, false claims can propagate through the agent network. We model this process as delayed consensus on a graph with grounded corrector nodes. Spectral decomposition by the grounded Laplacian yields a closed-form stability threshold for the verification dose: correction that is too strong or too delayed can turn consensus into oscillation. The most unstable regime occurs when the communication and verification delays coincide; for delay two, the threshold is the inverse golden ratio. The same framework gives a supermodular placement objective and a greedy (1-1/e)-approximation rule for assigning a limited corrector budget to influential nodes. Experiments across five open models confirm the predicted dose-delay oscillations. By contrast, grounded factual answering makes truth an absorbing boundary and eliminates the effect, suggesting that the instability is specific to signed-belief tasks while grounded verification remains stabilizing

---

### Towards Evaluation of Implicit Software World Models in Coding LLMs

Authors: Egor Bogomolov, Yaroslav Zharov
Published: 2026-06-25
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2606.27406v1

Software engineering, whether performed by humans or by AI agents, requires reasoning about how software behaves. We call the internal model that supports such reasoning the software world model, and view current code-execution benchmarks as covering one well-studied slice of it -- control flow. In this paper, we take a step toward a broader evaluation by shifting the observable axis to execution resources: alongside test outcome and exception class, we predict peak memory, wall-clock time, and ranked profiler outputs at method and line granularity. We use SWE-bench Verified as the source of data to hold the test close to real-world software engineering tasks. All tested models, frontier ones included, show modest performance and brittle behaviour, suggesting a notable lack of understanding of how software is executed, as opposed to how its source code is written.

---

### Automated brain tumor detection in MRI images using CNN and ResNet architectures

Authors: Annapurna V K, Asha N, K Paramesha, Shabana Sultana, Kirankumar Humse
Published: 2026-06-25
Categories: eess.IV, cs.AI
arXiv: https://arxiv.org/abs/2606.27405v1

Deep learning has shown significant potential in medical image analysis, particularly for disease detection using MRI scans. Accurate and early diagnosis of brain tumors remains challenging due to the complexity of brain structures and reliance on manual interpretation. This work presents an automated deep learning-based approach for brain tumor detection from MRI images using Convolutional Neural Networks and Residual Networks. Transfer learning is applied with two pretrained architectures, ResNet18 and ResNet50, to classify MRI scans into tumor and non-tumor categories. Experiments are conducted on a dataset of 3,929 brain MRI images, evaluating the impact of model depth and fine-tuning strategies. The results show that ResNet18 achieves a higher accuracy of 97% compared to 96% for ResNet50, demonstrating better generalization on limited medical data. The proposed framework enables fast, accurate, and cost-effective brain tumor detection, supporting early diagnosis and clinical decision-making.

---

### DanceOPD: On-Policy Generative Field Distillation

Authors: Wei Zhou, Xiongwei Zhu, Zelin Xu, Bo Dong, Lixue Gong, Yongyuan Liang, Meng Chu, Leigang Qu, Lingdong Kong, Wei Liu, Tat-Seng Chua
Published: 2026-06-25
Categories: cs.CV, cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2606.27377v3

Modern image generation demands a single model that unifies diverse capabilities, including text-to-image (T2I), local editing, and global editing. However, these capabilities are rarely naturally aligned and often conflict. For instance, editing tends to degrade T2I performance, while global and local editing interfere with each other. Consequently, effectively composing these capabilities has become a central challenge for image generation model training. To tackle this, we introduce DanceOPD, an on-policy generative field distillation framework for flow-matching models that routes each sample to one capability field, queries one low-noise student-induced state, and trains with a simple velocity MSE objective. With each capability source defined as a velocity field over the shared flow state space, the student learns from fields queried on its own rollout states to compose expert capabilities. This formulation also absorbs operator-defined fields such as classifier-free guidance. Comprehensive experiments on T2I, editing, realism-field absorption, and CFG absorption show that our approach improves multi-capability composition, strengthening target capabilities while preserving anchor generation quality. We believe this work establishes a practical route for generative field distillation in flow-matching models.

---

### DanceOPD: On-Policy Generative Field Distillation

Authors: Wei Zhou, Xiongwei Zhu, Zelin Xu, Bo Dong, Lixue Gong, Yongyuan Liang, Meng Chu, Leigang Qu, Lingdong Kong, Wei Liu, Tat-Seng Chua
Published: 2026-06-25
Categories: cs.CV, cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2606.27377v2

Modern image generation demands a single model that unifies diverse capabilities, including text-to-image (T2I), local editing, and global editing. However, these capabilities are rarely naturally aligned and often conflict. For instance, editing tends to degrade T2I performance, while global and local editing interfere with each other. Consequently, effectively composing these capabilities has become a central challenge for image generation model training. To tackle this, we introduce DanceOPD, an on-policy generative field distillation framework for flow-matching models that routes each sample to one capability field, queries one low-noise student-induced state, and trains with a simple velocity MSE objective. With each capability source defined as a velocity field over the shared flow state space, the student learns from fields queried on its own rollout states to compose expert capabilities. This formulation also absorbs operator-defined fields such as classifier-free guidance. Comprehensive experiments on T2I, editing, realism-field absorption, and CFG absorption show that our approach improves multi-capability composition, strengthening target capabilities while preserving anchor generation quality. We believe this work establishes a practical route for generative field distillation in flow-matching models.

---

### Ask, Solve, Generate: Self-Evolving Unified Multimodal Understanding and Generation via Self-Consistency Rewards

Authors: Ritesh Thawkar, Shravan Venkatraman, Omkar Thawakar, Abdelrahman Shaker, Fahad Khan, Hisham Cholakkal, Salman Khan, Rao Muhammad Anwer
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27376v1

Most unified large multimodal models (LMMs) that support both visual understanding and image generation still rely on curated post-training supervision, such as human annotations, preference labels, or external reward models. We ask whether a unified LMM can improve both abilities autonomously using only unlabeled images. We propose a self-evolving training framework with three internal roles: a Proposer that generates visual questions, a Solver that answers and evaluates them, and a Generator that synthesizes images. Training uses only self-derived consistency signals, without human annotations, preference labels, or task-trained external reward/judge models. To stabilize learning, we introduce Solver Token Entropy (STE), a continuous difficulty signal based on token-level prediction uncertainty that remains useful even when sample-level consistency becomes unreliable. For image generation, we design a multi-scale internal evaluation scheme that combines question-answer fidelity scoring with cycle-consistent captioning. This creates a solver-mediated coupling, where better visual understanding enables more reliable generation assessment and stronger internal training signals. The framework preserves the same role decomposition, reward logic, and training schedule across diffusion-based BLIP3o, rectified-flow BAGEL, and autoregressive VARGPT-v1.1 architectures, requiring only each backbone's native prompting and generation interface. Across eight understanding metrics, our method consistently improves over the corresponding base models. On BAGEL, it achieves a $+3.5\%$ absolute gain on MMMU and improves GenEval image generation performance from $82\%$ to $85\%$. Code and models are publicly released.

---

### World Action Models Enable Continual Imitation Learning with Recurrent Generative Replays

Authors: Manish Kumar Govind, Dominick Reilly, Smit Patel, Hieu Le, Srijan Das
Published: 2026-06-25
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2606.27374v1

Going beyond predicting robot actions, World Action Models (WAMs) can also generate future visual observations. We build on this generative capability to propose Recurrent Generative Replay (REGEN), a continual imitation learning framework that synthesizes pseudo-replay trajectories, enabling a robot policy to rehearse previously learned tasks without storing their original human demonstrations. During continual adaptation, REGEN recursively queries the WAM to synthesize pseudo-replay trajectories conditioned only on prior task instructions and current-task observations. Experiments in both simulation and real-world manipulation settings show that REGEN reduces catastrophic forgetting by up to $50\%$ relative to sequential fine-tuning, while approaching the performance of privileged experience replay methods that require access to real replay data. Finally, we analyze the factors limiting generated replay, identifying long-horizon visual degradation and action-observation inconsistency as the primary bottlenecks. Our results establish WAMs as a promising foundation for continual robot learning without stored demonstrations.

---

### Paying More Attention to Visual Tokens in Self-Evolving Large Multimodal Models

Authors: Shravan Venkatraman, Ritesh Thawkar, Omkar Thawakar, Rao Muhammad Anwer, Hisham Cholakkal, Salman Khan, Fahad Khan
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27373v1

Recently, self-evolving large multimodal models (LMMs) have received attention for improving visual reasoning in a purely unsupervised setting. However, multi-role self-play and self-consistency reward schemes in existing self-evolving LMMs optimize answer agreement without ensuring the decoder attends to visual content, relying instead on statistical language priors to produce self consistent outputs. This leads to a persistent failure mode we term visual under-conditioning, where the decoder relies on language priors rather than the image during generation, manifesting as insufficient attention to visual tokens. As a result, current self-evolving LMMs struggle on vision--language understanding tasks such as image captioning and visual question answering. To address this, we propose VISE (Visual Invariance Self-Evolution), a purely unsupervised self-evolving framework that directly regularizes the model's visual conditioning policy through two complementary invariance-based rewards: a geometric invariance reward that enforces spatial consistency under known transformations, and a semantic invariance reward that penalizes evidence-agnostic generation by requiring the model to recognize the absence of evidence when predicted regions are perturbed. VISE operates within a single model without specialist roles, external reward models, or annotations, and is trained on raw unlabeled images. Experiments on 18 benchmarks demonstrate the efficacy of our approach. Using Qwen3-VL-2B as the base model, VISE achieves gains of $+16.85$ CIDEr on COCO and $+19.66$ CIDEr on TextCaps, reduces object hallucination by $5.0$ Chair-I points, and generalizes across four model families and scales. Our code and models are available at https://mbzuai-oryx.github.io/VISE

---

### DnA: Denoising Attention for Visual Tasks

Authors: Ron Campos, Subhajit Maity, Xin Li, Srijan Das, Aritra Dutta
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27372v2

The softmax activation in multihead attention (MHA) is the de facto standard for attention-based models in visual perception tasks. However, standard softmax can produce noisy attention patterns that dilute relevant features and degrade its performance. In this paper, we propose Denoising Attention or DnA, in which, first, a positive query identifies which image features belong to the correct class, and a negative query identifies closely associated but irrelevant image features. DnA then projects these interactions into two distinct subspaces with larger principal angles, promoting subspace separation and improved discriminability. Using a ViT-B backbone, our proposed DnA achieves an absolute gain of 0.8% on ImageNet-1K compared to the baseline. We further show improvements across multiple visual understanding tasks, including video understanding with video transformers (1.8%) and video LLMs (0.5%). Our extensive empirical analyses justify the design choices involving two interacting subspaces and the denoising effect of DnA.

---

### DnA: Denoising Attention for Visual Tasks

Authors: Ron Campos, Subhajit Maity, Xin Li, Srijan Das, Aritra Dutta
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27372v1

The softmax activation in multihead attention (MHA) is the de facto standard for attention-based models in visual perception tasks. However, standard softmax can produce noisy attention patterns that dilute relevant features and degrade its performance. In this paper, we propose Denoising Attention or DnA, in which, first, a positive query identifies which image features belong to the correct class, and a negative query identifies closely associated but irrelevant image features. DnA then projects these interactions into two distinct subspaces with larger principal angles, promoting subspace separation and improved discriminability. Using a ViT-B backbone, our proposed DnA achieves an absolute gain of 0.8% on ImageNet-1K compared to the baseline. We further show improvements across multiple visual understanding tasks, including video understanding with video transformers (1.8%) and video LLMs (0.5%). Our extensive empirical analyses justify the design choices involving two interacting subspaces and the denoising effect of DnA.

---

### Don't Settle at the Mode! Mitigating Diversity Collapse in Pretrained Flow Models via Feature Self-Guidance

Authors: Pradhaan S Bhat, Rishubh Parihar, Abhijnya Bhat, R. Venkatesh Babu
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27371v1

State-of-the-art flow models generate stunning images from text or image prompts. However, they suffer from diversity collapse when generating multiple samples under the same conditioning. Existing methods address this issue via either latent guidance, which has limited effectiveness, or sample selection, which relies on external reward models that incur significant inference-time overhead. In this work, we introduce an efficient, training-free self-guidance mechanism to mitigate diversity collapse without requiring additional reward models. Specifically, we disperse the internal features of the flow model during batch generation with feature self-guidance. Further, to keep the features close to the manifold, we introduce a manifold regularization step that projects these dispersed features back onto the data manifold, ensuring diverse generation without sacrificing alignment with the input conditions. Our method integrates seamlessly as a plug-and-play module into pretrained flow models, adding only a marginal inference cost. Experiments demonstrate significant improvements in diversity while preserving fidelity across several conditional flow models, including multi-step and few-step text-to-image, depth-to-image, and reference image generation.

---

### Reinforcement Learning without Ground-Truth Solutions can Improve LLMs

Authors: Yingyu Lin, Qiyue Gao, Nikki Lijing Kuang, Xunpeng Huang, Kun Zhou, Tongtong Liang, Zhewei Yao, Yi-An Ma, Yuxiong He
Published: 2026-06-25
Categories: cs.LG
arXiv: https://arxiv.org/abs/2606.27369v1

Reinforcement learning with verifiable rewards (RLVR) for training LLMs typically rely on ground-truth answers to assign rewards, limiting their applicability to tasks where the ground-truth solution is unknown. We introduce a \textbf{R}anking-\textbf{i}nduced \textbf{VER}ifiable framework (RiVER) that trains LLMs on score-based optimization tasks without ground-truth solutions, using deterministic execution feedback as continuous-valued supervision. When applying group-relative RL to such continuous rewards, we identify two key challenges: \emph{scale dominance}, where uncalibrated score magnitudes across test instances distort policy updates, and \emph{frequency dominance}, where repeatedly sampled suboptimal solutions can outweigh rare but stronger candidates. RiVER addresses these challenges with calibrated reward shaping that uses instance-wise comparisons and emphasizes top-ranked solvers while retaining bounded feedback for other valid solutions. We train on 12 AtCoder Heuristic Contest tasks and evaluate on Algorithm Engineering Benchmark (ALE-Bench), LiveCodeBench, and USACO. RiVER advances Qwen3-8B and GLM-Z1-9B-0414 by 8.9\% and 9.4\% in ALE rating rank. More importantly, despite training exclusively on score-based tasks without any ground-truth solutions, RiVER also improves the backbones across exact-solution benchmarks such as LiveCodeBench and USACO by an absolute average improvement of 2.4\% and 3.5\%. By contrast, baselines trained with raw execution scores improve ALE rating but fail to transfer to exact-solution benchmarks. These results suggest that score-based optimization tasks, combined with proper reward calibration, can serve as effective training environments for general coding ability without ground-truth solutions.

---

### PhysiFormer: Learning to Simulate Mechanics in World Space

Authors: Yiming Chen, Yushi Lan, Andrea Vedaldi
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27364v1

We present PhysiFormer, a diffusion transformer for physically-plausible 3D object motion. Unlike video world models that operate in view-dependent pixel space, PhysiFormer represents objects as 3D meshes expressed in world coordinates. Given the initial vertex positions and velocities, as well as object material type, rigid or elastic, the model samples future vertex trajectories. While related neural physics approaches build on ad-hoc latent spaces or explicitly enforce rigidity and causality, PhysiFormer shows that excellent results can be obtained without any such inductive biases, by casting vertex trajectory prediction as a single denoising diffusion process directly in world coordinates. The probabilistic formulation captures uncertainty in the learned dynamics, enabling diverse plausible futures from initial conditions, making this framework potentially useful for applications with unobserved uncertainty. The model features attention factorised over time, space, and objects for efficiency, enabling permutation-invariant multi-object reasoning without needing explicit object encoding. Trained on over 100k simulated trajectories, PhysiFormer generates rigid and elastic mechanics, and generalises to mixed-material settings, unseen real-world geometries, and larger object counts. It substantially outperforms autoregressive baselines in trajectory accuracy, rigidity preservation, and momentum-based physical consistency. Our results position coordinate-space diffusion as a promising step toward view-invariant, geometry-aware world modelling for robotics, graphics, and physical design. Visualisations, code, and models are available at https://yimingc9.github.io/physiformer.

---

### Autoregressive Boltzmann Generators

Authors: Danyal Rehman, Charlie B. Tan, Yoshua Bengio, Avishek Joey Bose, Alexander Tong
Published: 2026-06-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2606.27361v1

Efficient sampling of molecular systems at thermodynamic equilibrium is a hallmark challenge in statistical physics. This challenge has driven the development of Boltzmann Generators (BGs), which allow rapid generation of uncorrelated equilibrium samples by combining a generative model with exact likelihoods and an importance sampling correction. However, modern BGs predominantly rely on normalizing flows (NFs), which either suffer from limited expressivity due to strict invertibility constraints (discrete time) or computationally expensive likelihoods (continuous time). In this paper, we propose Autoregressive Boltzmann Generators (ArBG) -- a novel autoregressive modelling framework -- that overcomes these limitations by departing from the flow-based BG paradigm. ArBG circumvents the topological constraints of flows and enables sequential inference-time interventions, while offering enhanced scalability by leveraging architectures effective in Large Language Models. We empirically demonstrate that ArBG leads to significant improvements over flow-based models across all benchmarks, but particularly in larger peptide systems such as the 10-residue Chignolin. Furthermore, we introduce Robin, a 132 million parameter transferable model trained with the ArBG framework which improves over the previous state-of-the-art, reducing the zero-shot energy error, E-W$_2$, on 8-residue systems by over 60$\%$. The code can be found at the following link: https://github.com/danyalrehman/autobg.

---

### When are likely answers right? On Sequence Probability and Correctness in LLMs

Authors: Johannes Zenn, Jonas Geiping
Published: 2026-06-25
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2606.27359v1

Many decoding methods for large language models can be understood as shifting probability mass toward outputs that are more likely under the model, either locally at the token level or globally at the sequence level. Therefore, their success depends on a fundamental question: when does sequence probability, that is, the conditional probability of a continuation given a prompt, actually align with correctness? In this paper, we set out to quantify this relationship across decoding methods, models, and benchmarks at four levels: across decoding methods, across hyperparameters within a method, across prompt-answer pairs within a dataset, and across repeated responses to the same prompt. We find that higher sequence probability is often predictive of correctness across prompt-answer pairs within a fixed dataset. However, this relationship does not generally transfer to decoding decisions: increasing sequence probability by changing hyperparameters or methods does not reliably improve accuracy. Further, sequence probability is not a good indicator of correctness for responses to the same prompt. These findings clarify when decoding can and cannot be expected to improve correctness, and provide practical guidance for decoding, self-consistency, and verifier-free self-improvement.

---

### Error-Conditioned Neural Solvers

Authors: Haina Jiang, Liam Wang, Peng-Chen Chen, Min Seop Kwak, Seungryong Kim, Brian Bell, Jeong Joon Park
Published: 2026-06-25
Categories: cs.LG, cs.AI, cs.CV, math.NA
arXiv: https://arxiv.org/abs/2606.27354v1

Neural surrogate models offer fast approximate mappings from PDE parameters to solutions, but they typically treat solving as a purely statistical task: once trained, they struggle to correct their own constraint violations and extrapolate beyond the training distribution. Recent hybrid methods promote physical correctness by targeting the PDE residual via gradient descent or Gauss--Newton steps, but inherit the compute cost and instability of the underlying classical optimizers. We show, theoretically and empirically, that numerically minimizing the PDE residual can be an unreliable proxy for reconstruction accuracy in ill-conditioned systems, explaining why these methods often do not make accurate predictions despite achieving low residuals. We propose error-conditioned Neural Solvers (ENS), built on a different principle: rather than an optimization target, the PDE residual field is passed as a direct input to the network at each iteration, enabling it to read the spatial structure of its own errors and learn an update policy to iteratively correct its predictions. Across four PDE families, ENS attains the highest prediction accuracy in the large majority of settings, with gains reaching $10\times$ on turbulent Kolmogorov flow, while avoiding the expensive compute cost of hybrid methods. ENS's learned correction policy generalizes under distribution shift, including zero-shot parameter changes and cross-equation transfer, where its relative advantage is largest in the ill-conditioned regimes where residual minimization is least reliable. Project website: https://neuralsolver.github.io/.

---

### All you need is log

Authors: Akshay Balsubramani
Published: 2026-06-25
Categories: cs.IT, math.PR, math.ST, stat.ML
arXiv: https://arxiv.org/abs/2606.27349v3

How different are several probability distributions from one another? For two distributions the standard answer is the family of Rényi divergences, singled out by two natural requirements: processing the data never makes distributions easier to tell apart, and independent repetitions add. Many problems in learning and statistics compare more than two distributions at once, such as testing among several hypotheses or bounding generalization against several priors. The same two requirements leave one kind of building block, built on a coincidence probability: how unlikely it is that independent samples, one from each distribution, all show the same empirical distribution. The logarithm is forced because repetitions add, which is already visible for a single experiment repeated. This characterization is known in greater generality, and this paper is about the meaning of its building blocks. On a finite alphabet, each building block indexed by a rational point of the simplex is the exponential rate of that coincidence as the samples grow in fixed proportions. Each is also the limiting free energy of Bayesian inference over distributions. At any amount of data, the free energy of the posterior is the coincidence measure plus two costs: the expected distance from a posterior draw to the most likely distribution, and the information gained per unit of data. Both costs vanish as data accumulate. When the comparison is conditioned on side information, every kind of building block has a conditional counterpart, and the coincidence ones alone do not suffice.

---

### All you need is log

Authors: Akshay Balsubramani
Published: 2026-06-25
Categories: cs.IT, math.PR, math.ST, stat.ML
arXiv: https://arxiv.org/abs/2606.27349v2

Comparing two probability distributions is a basic building block of statistics and machine learning, and the right family is well understood: the Rényi divergences of order $α\in(0,\infty)$ are the unique family monotone under data processing and additive on independent products. Many problems instead compare more than two distributions at once -- multi-population fairness, multi-prior PAC-Bayes bounds, multi-hypothesis testing -- and the right multi-distribution generalization of the Rényi family has been an open question. We characterize it. Every functional of $W$-tuples of distributions that is monotone under data processing and additive on independent products is a positive integral of multi-way coincidence divergences $C_α(π_1,\dots,π_W) := -\log\int π_1^{α_1}\cdotsπ_W^{α_W}$ (with $\sum_k α_k = 1$) over a parameter space with four strata: the simplex interior; mixed-sign exponent cones (the analogue of Rényi orders $>1$); a tropical boundary at infinity carrying max-divergences; and pairwise Kullback-Leibler edges at the simplex vertices. Each stratum is necessary -- the destination of an explicit data-processing-monotone, product-additive divergence the others cannot reproduce -- and each is a clean limit of simplex-interior atoms. The same family arises from several independent routes -- the structural axioms, Kolmogorov-Nagumo means with Rényi's entropy axiomatics, classical entropy characterizations, multi-hypothesis testing error exponents, and a multi-lottery betting interpretation -- structural evidence that this is the canonical multi-distribution Rényi calculus rather than an artefact of any one axiomatic input. The two-prior case recovers the standard Rényi result; a worked $W=3$ instance, numerical verification, and a conditional extension round out the treatment.

---

### Mapping Political-Elite Networks in Europe with a Multilingual Joint Entity-Relation Extraction Pipeline

Authors: Kirill Solovev, Jana Lasser
Published: 2026-06-25
Categories: cs.CL
arXiv: https://arxiv.org/abs/2606.27347v3

Whether political elites organise into rent-seeking coalitions that capture public resources or civic networks that sustain governance is a central question in comparative politics. Yet observing these complex, informal, and adversarial ties at scale has historically required intensive manual coding, while automated text-as-data methods have largely been limited to simple co-occurrence. Recent large language model (LLM) approaches offer a path forward but often rely on proprietary APIs, lack cross-lingual capability, and struggle with scalable entity resolution. We present a modular, fully open-weight pipeline for multilingual joint entity-relation extraction that builds signed, temporal knowledge graphs from massive unstructured news corpora. It combines span-based named-entity recognition (NER) with a three-stage linking cascade mapping mentions to language-independent Wikidata identifiers; a high-throughput, ontology-constrained mixture-of-experts model then uses guided decoding to extract directed, signed relationships grounded in a domain ontology. A full-coverage spot-check against a 3491-relation gold standard shows high textual correctness (68.2% strict to 93.7% lenient). Two large-scale case studies validate the pipeline against the public record. In Austria, it reconstructs a political party's complete lifecycle, dating internal fractures and tracking personnel into successor factions and court convictions. In a Polish corpus, it uncovers the overlapping economic and governance networks of state-enterprise patronage, alongside the structurally balanced, signed conflict network of the polarized Civic Platform (Platforma Obywatelska, PO)--Law and Justice (Prawo i Sprawiedliwość, PiS) duopoly. By bridging raw multilingual text and structured relational data, our framework provides a robust, replicable foundation for cross-national empirical computational social science.

---

### SCoPE: Sightline-Coordinate Positional Encoding for Video Diffusion Transformers

Authors: Minghao Yin, Jiahao Lu, Wenbo Hu, Wang Zhao, Shan Ying, Kai Han
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27345v3

Video diffusion transformers address their tokens by position on the pixel-time grid: an address in the tensor, not in the world. The address we would want, the world point a token depicts, lies on a surface not yet generated, while its camera ray is fixed once the user specifies a trajectory. SCoPE therefore treats the ray as a second positional coordinate, and camera control becomes a property of the coordinate system, not an added module. The ray is added to the pretrained attention's queries and keys, and the score gains a term that reads the two rays alone. Its canonical form, the reciprocal product of line geometry, measures how nearly two lines of sight meet. Normalize-Gate-Inject makes a single encoding trainable across metric and up-to-scale pose sources. The retrofit keeps RoPE bit-exact, starts from the unchanged pretrained DiT, and adds under 0.1/% new parameters. On Wan2.2 at 5B and 14B under matched data and budget, SCoPE improves every camera-controllability and fidelity metric, leads all closed-loop revisit metrics, and shows widening margins with model size. At 14B, rotation error falls 29/% and FVD 43/% below the strongest baseline.

---

### RayPE: Ray-Space Positional Encoding for 3D-Aware Video Generation

Authors: Minghao Yin, Jiahao Lu, Wenbo Hu, Wang Zhao, Shan Ying, Kai Han
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27345v2

Modern video diffusion transformers position their tokens through RoPE on the (u,v,t) axes -- a description of the camera's sampling grid that says nothing about the 3D structure of the scene. We observe that the geometric relation between two camera rays is captured by the Plucker reciprocal product, which is bilinear in the two rays -- the same algebraic form as the dot product in Transformer attention. Building on this analogy, we propose RayPE, a positional-encoding extension that injects per-token 6D Plucker coordinates additively into the queries and keys of self-attention, with a query/key flip arrangement under which the symmetric identity configuration coincides exactly with the reciprocal product. The injection is additive, the resulting attention score decomposes into a content term, a geometry term, and two content and geometry cross-terms -- all of which our experiments find individually necessary. To make the encoding stable across video data with heterogeneous camera-translation scales (SfM, deep SLAM, metric), we further decouple ray direction from moment magnitude, gate the encoding by a learned function of the log-magnitude, and apply RMSNorm to align it with the QKNorm-normalized content branch. The full module adds less than 0.1% parameters to a pretrained video DiT, is zero-initialized to start from the pretrained weights, and improves camera controllability, cross-frame 3D consistency, and overall video quality on a four-dataset training mixture.

---

### Understanding Domain-Aware Distribution Alignment in Budgeted Entity Matching

Authors: Nicholas Pulsone, Gregory Goren, Roee Shraga
Published: 2026-06-25
Categories: cs.DB, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2606.27342v2

Entity Matching (EM) is a core operation in the data integration pipeline, where records from different sources are compared to determine whether they refer to the same real-world entity. Recent work has incorporated domain information and low-resource learning techniques to better adapt EM systems to realistic settings. While these approaches have demonstrated strong performance, it remains unclear how they behave under varying data constraints and levels of supervision in practice. In this paper, we investigate a state-of-the-art method for low-resource, domain-aware EM--BEACON--and study how its performance is affected by different algorithmic choices and data availability conditions. We conduct a series of targeted experiments to evaluate these variations, providing deeper insight into the role of distribution alignment and the behavior of the BEACON framework.

---

### SAM2Matting: Generalized Image and Video Matting

Authors: Ruiqi Shen, Guangquan Jie, Chang Liu, Henghui Ding
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27339v1

Despite impressive advances in image matting, video matting remains challenging due to the inherent gap between high-level tracking, which requires frame-wise understanding, and low-level matting, which focuses on extremely fine-grained details. Existing methods attempt this with expensive and narrowly-scoped video matting datasets, which may limit out-of-domain generalization and compromise tracking robustness. We rethink the paradigm with SAM2Matting, a tracker-to-matting framework that advances VOS trackers to high-fidelity video matting. Specifically, it decouples the task by enhancing a foundational tracker (e.g., SAM2, SAM3) with a region-proposal bridge and dedicated matting heads, enabling the uncompromised tracker to handle temporal consistency while the matting components resolve fine-grained details. Notably, despite being trained only on images, SAM2Matting establishes new state-of-the-art performance on video matting, supports diverse prompt types, maintains strong temporal consistency, and demonstrates robust generalization across both human-centric and in-the-wild scenarios.

---

### Language-Based Digital Twins for Elderly Cognitive Assistance

Authors: Mohammad Mehdi Hosseini, Mohammad H. Mahoor, Hiroko H. Dodge
Published: 2026-06-25
Categories: cs.AI
arXiv: https://arxiv.org/abs/2606.27334v1

Digital twins have emerged as a promising paradigm for personalized healthcare, enabling modeling of individual behavior and health trajectories. In cognitive health, early detection of Mild Cognitive Impairment (MCI) remains challenging, where language and conversational patterns serve as non-invasive biomarkers. In this work, we propose a language-based digital twin framework that leverages large language models (LLMs) to mimic the conversational behavior of elderly individuals by incorporating stylometric cues and contextual metadata. To evaluate fidelity and cognitive consistency, we introduce a multi-head conditional variational autoencoder (cVAE) that jointly measures reconstruction quality and predicts cognitive scores. Experiments on the I-CONECT dataset show that the digital twin preserves identity-specific characteristics and achieves reconstruction and MoCA prediction errors comparable to real data, while outperforming baseline GPT-generated responses. These results highlight the potential of language-based digital twins as a scalable and non-invasive approach for personalized and continuous cognitive health monitoring.

---

### RoPEMover: Depth-Aware Object Relocation via Positional Embeddings

Authors: Ipek Oztas, Duygu Ceylan, Aybars Bugra Aksoy, Aysegul Dundar
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27332v1

Moving an object in a single image requires geometry-consistent spatial rearrangement, including handling occlusions, revealing previously unseen regions, and maintaining coherent shadows and reflections. Existing approaches are not well suited to this setting and often fail to preserve such scene-level consistency. We address this problem by introducing a geometry-aware object motion method that operates directly on the positional representations of diffusion transformers. Our key insight is that rotary positional embeddings (RoPE) define a structured spatial field that can be explicitly manipulated to induce controlled motion. We extend 2D RoPE into a depth-aware formulation that encodes 3D spatial structure, enabling consistent object displacement and scene-aware updates. Our model is trained using synthetic data combined with a small set of real images via parameter-efficient fine-tuning. Despite minimal real supervision, it preserves object identity under large spatial displacements, generates plausible content in newly revealed regions, and consistently updates scene-dependent effects such as shadows and illumination. Experimental results on standard object motion benchmarks demonstrate state-of-the-art performance across all evaluation metrics.

---

### Empowering GUI Agents via Autonomous Experience Exploration and Hindsight Experience Utilization for Task Planning

Authors: Tianyi Men, Zhuoran Jin, Pengfei Cao, Yubo Chen, Kang Liu, Jun Zhao
Published: 2026-06-25
Categories: cs.CL, cs.AI, cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2606.27330v1

Multimodal web agents can assist humans in operating repetitive GUI tasks, where effective task planning is essential for decomposing complex tasks into executable actions. While small open source MLLMs are cost efficient and privacy preserving compared with commercial large models, they suffer from weak planning and limited cross website generalization. To address these limitations, we introduce the planning experience exploration and utilization (PEEU) method, which autonomously explores environments to discover experiences and utilizes hindsight experience to synthesize strictly aligned, high level training data. To quantitatively analyze the generalization behaviors driving this performance, we propose the task decomposition hierarchical analysis framework (TDHAF) to systematically study compositional generalization across three task granularities: low, middle and high levels. Our analysis reveals that mastering low level atomic skills does not guarantee high level planning competence, while high level task training yields stronger OOD generalization. Experiments on real world benchmarks demonstrate PEEU's superior effectiveness: our 7B model achieves 30.6% accuracy, outperforming the much larger Qwen2.5-VL-32B model. These demonstrate constructing hindsight high level tasks and leveraging experiences is crucial for OOD planning abilities of small MLLMs.

---

### Hallucination in World Models is Predictable and Preventable

Authors: Nicklas Hansen, Xiaolong Wang
Published: 2026-06-25
Categories: cs.LG, cs.CV, cs.RO
arXiv: https://arxiv.org/abs/2606.27326v1

Modern generative world models render increasingly realistic action-controllable futures, yet they frequently hallucinate: rollouts remain visually fluent while drifting from the ground-truth dynamics. We hypothesize that hallucination concentrates in low-coverage regions of the state-action space, where lightweight data-centric signals can both detect it and guide mitigation. To test this, we introduce MMBench2, a 427-hour, 210-task dataset for visual world modeling with ground-truth actions, rewards, and live simulators, and train a 350M-parameter world model on it. We identify three distinct hallucination modes: perceptual, action-marginalized, and scene-diverging -- each anchored to a different stage of the pipeline, and develop three signals that accurately predict where the model will fail. To close coverage gaps at training time, we develop a coverage-aware sampling technique; to close them online, our hallucination predictors serve as curiosity rewards for targeted data collection, yielding a data-efficient finetuning recipe that adapts the pretrained world model to entirely unseen environments with as few as 50 real environment trajectories. Overall, our findings reveal that hallucination in world models is inherently a data coverage issue, and that the same signals used to detect it can also be used for mitigation. An interactive web version of our paper is available at https://www.nicklashansen.com/mmbench2

---

### Not All Actions Are Equal: Rethinking Conditioning for Dexterous World Model

Authors: Zizhao Yuan, Zhengtu Liang, Taowen Wang, Qiwei Liang, Yichi Wang, Yunheng Wang, Yuetong Fang, Lusong Li, Zecui Zeng, Renjing Xu
Published: 2026-06-25
Categories: cs.CV
arXiv: https://arxiv.org/abs/2606.27325v1

Recent advances in action-conditioned world models show promising progress in modeling complex interactions and forecasting future states under diverse action sequences. While these models are often driven by stronger visual representations and model capacity, action conditioning itself remains underexplored. Most existing approaches compress the entire action sequence into a single representation, which works well for low-DoF control but becomes less reliable in high-DoF scenarios. We observe that high-DoF dexterous actions are inherently heterogeneous, spanning multiple orders of magnitude, where large-scale motions coexist with subtle but important signals. When uniformly aggregated, optimization exhibits an imbalance across action components, which hinders the modeling of fine-grained effects and affects action fidelity. We therefore propose DexAC-WM, which treats action conditioning as a structured process rather than global compression. DexAC preserves dimension-level semantics via action tokenization and aligns action signals with visual dynamics through local refinement and global modulation. To address the limited high-level semantic grounding in existing world models, we further introduce a semantic branch that provides rich object-scene priors, which enables world model to capture dynamic visual details while supporting high-DoF action-conditioned video prediction. Experiments on EgoDex and EgoVerse show that combining the semantic branch with DexAC significantly improves FID, FVD, and PCK, demonstrating gains in visual-temporal realism and action-following consistency. We further verify that DexAC extends to other backbones, showing the scalability of our structured action-conditioning design. These results suggest that scaling world models to high-DoF control requires both structured action modeling and semantic grounding.

---

### Beyond the Hard Budget: Sparsity Regularizers for More Interpretable Top-k Sparse Autoencoders

Authors: Nathanaël Jacquier, Maria Vakalopoulou, Mahdi S. Hosseini
Published: 2026-06-25
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2606.27321v2

Sparse autoencoders (SAEs) have become a leading tool for interpreting the representations of vision foundation models, decomposing their polysemantic activations into a larger set of sparse, more monosemantic features. The Top-$k$ SAE, a now-standard variant, enforces sparsity architecturally through its activation function, retaining only the $k$ most active latents per input. Because it was designed precisely to avoid the $\ell_1$ penalty used by earlier SAEs and its known drawbacks, it has not been combined with an explicit sparsity regularizer. Yet the Top-$k$ SAE retains limitations of its own, and we hypothesize that a sparsity penalty acting before the selection could sharpen each latent's selectivity and make the code more interpretable, without reintroducing the drawbacks of the $\ell_1$ penalty. We introduce two sparsity regularizers compatible with the Top-$k$ architecture, both acting on the activations before the Top-$k$ selection: an $\ell_1$ penalty on the unselected (off-support) units, and a scale-invariant $\ell_1/\ell_2$-ratio penalty that concentrates the code onto fewer effective units. Both penalties are applied only to the batch-active units, those selected by the Top-$k$ operator at least once within the batch. Across two datasets, three vision foundation models, and a range of $k$, both regularizers consistently improve monosemanticity at no cost to reconstruction quality. The $\ell_1/\ell_2$ penalty further concentrates information into fewer latents, making reconstruction more robust to the inference-time choice of $k$ and improving small-budget linear probing. Our central finding is that hard architectural sparsity and soft sparsity regularization are complementary rather than mutually exclusive.

---

### Elastic Time: Dynamic Frame Rate Bottlenecks for Neural Audio Coding

Authors: Dimitrios Bralios, Paris Smaragdis, Minje Kim
Published: 2026-06-25
Categories: cs.SD, cs.LG, eess.AS
arXiv: https://arxiv.org/abs/2606.27320v1

Neural audio autoencoders have become a core component of compression, feature extraction, and generation. However, while existing systems support variable bitrate, the vast majority of models still operate at a fixed latent frame-rate, allocating equal temporal budget to regions with very different information density, which can result in unnecessarily long sequences. We introduce Elastic Time, a dynamic frame-rate bottleneck that converts fixed-frame-rate autoencoders to dynamic ones. Our method learns a lightweight latent predictor used to decide which frames can be skipped and later reconstructed, enabling efficient greedy boundary selection at inference. Experiments show our method enables deployment-time rate control while improving efficiency-quality tradeoffs relative to baselines. Overall, we provide a flexible mechanism for adjusting temporal resolution in audio autoencoders, potentially facilitating more efficient downstream modeling for generation and long-context tasks.

---

### OctoSense: Self-Supervised Learning for Multimodal Robot Perception

Authors: Anthony Bisulco, Jeremy Wang, Kostas Daniilidis, Randall Balestriero, Pratik Chaudhari
Published: 2026-06-25
Categories: cs.CV, cs.RO
arXiv: https://arxiv.org/abs/2606.27317v1

We present OctoSense, an open-source sensor platform with stereo RGB and event cameras, LiDAR, a thermal camera, an inertial measurement unit, RTK-corrected global positioning system, and proprioception (CAN bus data from a car, and joint angles for a quadruped robot). The eponymous OctoSense dataset contains 59 hours of time-synchronized driving data across different types of environments at different times of the day, including situations with highly degraded sensors. We demonstrate multi-modal self-supervised learning using such real-world robotics data, where sensors have different representations, frequencies, latencies and noise. Our approach, a "late-fusion" masked autoencoder, (i) uses modality-specific tokenizers to account for different spatiotemporal characteristics of these sensors, and (ii) caches modality-specific tokens at inference time to process new measurements as they come. This architecture (i) is fast (6.68 ms and 112 ms on NVIDIA 5090 and Orin NX respectively, to compute the representation), (ii) performs better than existing image-only foundation models on tasks such as estimation of optical flow, depth, semantic segmentation, and ego-motion (translation, rotation, and steering angle), and (iii) predicts robustly at nighttime or in situations where sensory data is degraded. See our project page for links to the dataset, code, and supplementary videos: https://abisulco.com/octosense/.

---