Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
High-fidelity solubility prediction is fundamental to pharmaceutical development and environmental partitioning, where accurate modeling must couple molecular structure with thermodynamic behavior acr...
Implicit neural representations (INRs) provide a flexible framework for modeling high-dimensional continuous fields, but their training is often inefficient due to uniform subsampling that ignores spa...
Urban environments are shaped by design choices with long-term implications for health, safety, and quality of life, yet evaluating proposed interventions remains costly, time-consuming, and often imp...
Existing image forgery detectors often suffer from generalization to unseen manipulation methods due to the limited ability to capture transferable forensic cues. Recent cross-reconstruction based met...
GraphCert introduces a method to bootstrap graph reasoning agents by generating graph‑grounded question‑answer pairs and certifying the supporting evidence. The approach uses a Bootstrapped Graph Quizzer to produce QA pairs, then executes and semantically curates the evidence into certified rubrics that guide reward‑based training of a Graph Solver. Experiments on five GRBENCH domains show GraphCert outperforms larger LLM agents and demonstrates robust policy transfer across heterogeneous graphs.
The paper introduces a transition-based derandomization framework for dense binary hypervector codebooks used in hyperdimensional computing. It targets two similarity families—exponential and linear decay with scalar separation—and separates the similarity law, derandomization variant, and generator construction. The authors formalize variants that constrain initial Hamming weight, update-count variability, and update balance, deriving exact finite-dimensional expressions for bias, variance, and RMS error, and validate the theory with simulations to guide practical codebook design.
The paper investigates the consistency of surrogate loss methods for classification and policy learning when the set of admissible classifiers is constrained, such as by interpretability or fairness requirements. It shows that hinge loss is the only surrogate that preserves consistency when constraints limit only the prediction set, but consistency can fail if constraints also restrict the functional form. The authors derive conditions guaranteeing consistency for hinge-risk-minimizing classifiers and use these results to design efficient hinge-loss-based procedures for monotone classification problems.
The paper introduces TSDA-Track, a Template-Search Domain Adaptation framework designed to reduce modality gaps in cross‑modal visual object tracking. Two variants are explored: Pre‑AFA TSDA‑Track uses adversarial alignment before transformer interaction, while Enc‑CFA TSDA‑Track applies contrastive alignment after interaction to strengthen cross‑modal correspondence. Experiments on datasets such as LasHeR, RGBT234, GTOT, and Anti‑UAV‑024 show that both variants outperform state‑of‑the‑art trackers, with Pre‑AFA achieving an SR/PR of 43.2/56.0 on RGBT234 under the modality‑switch protocol.
FoCLIP is a framework that creates adversarial examples to manipulate CLIP-based image quality metrics by reducing the alignment between image and text features. It uses stochastic gradient descent to combine feature alignment, score distribution balancing, and pixel‑guard regularization, enabling high CLIPscore predictions while maintaining visual fidelity. Experiments on artistic prompts and ImageNet show significant CLIPscore gains, and the authors also propose a color‑channel sensitivity detection method that achieves 91% accuracy.
Efficient Multi-Modal Planning with Reward-Guided Preference Optimization for Autonomous Driving proposes EMPlan, a hybrid trajectory planning method that combines sparse anchors with an offset refinement module for low-latency, high-accuracy predictions. The approach uses a two-stage training paradigm—pretraining followed by reward-guided fine-tuning—to improve safety without extra inference cost, leveraging rule-based reward signals and unpaired preference supervision. EMPlan is evaluated on the non-reactive NAVSIM benchmark, achieving a favorable balance between planning accuracy and efficiency under real-time constraints.
The paper introduces a reusable Semantic Web framework that aggregates fragmented evidence needed for Fundamental Rights Impact Assessments under the EU AI Act, focusing on high‑risk public sector categories such as employment and worker management and access to essential public services. A curated 150‑record corpus is annotated across four axes and serialized into a SPARQL‑queryable knowledge graph of 1,351 RDF triples, enabling five demonstration scenarios that retrieve 103 records (68.7% coverage). Evaluation against a 69‑record gold standard shows that LLM‑assisted classification in the employment domain yields a low κ of 0.045, highlighting challenges in automated fairness‑related evidence retrieval, while all artefacts are released openly for regulators, authorities, and SMEs.
The study investigates how internal representations of video diffusion models align with human visual cortex responses. It finds that representations used for future video generation in an autoregressive (AR) model better match cortical activity than those for observed video, with future‑generation alignment concentrated in higher‑order visual areas. A behavioral experiment further shows that humans prefer videos enhanced by layers that align more strongly with cortical responses.
SheafStain introduces a sheaf-theoretic Schr"odinger Bridge framework to improve virtual staining of gigapixel whole-slide images. By treating Vision Foundation Model features as sheaf-like sections, it integrates class and patch tokens to enforce spatial and biological coherence, mitigating patch-boundary artifacts. The method is evaluated on HER2, ER, PR, and Ki‑67 stains, outperforming six prior approaches on stitched 1024 × 1024 outputs.
The paper introduces LogiC-Diff, a logic-conditioned bi-stage diffusion framework that embeds Signal Temporal Logic (STL) specifications into AI-enabled cyber‑physical system (CPS) forecasting models. By using STL as a conditioning signal, the method repairs inputs and refines outputs to jointly mitigate adversarial perturbations and enforce desired temporal behaviors. Experiments on two real‑world CPS datasets show that LogiC-Diff consistently improves robustness and specification compliance across various sensor faults and cyber attacks, outperforming reconstruction‑based defenses.
Steering Fields introduce adaptive vector fields that re-estimate steering directions at each step of a flow-based text-to-image generation process, replacing the fixed global steering vectors traditionally used. By operating on noisy states, they provide a continuous trade-off between steering strength and content preservation, and allow simultaneous induction and inhibition of concepts without explicit spatial masks or object priors. The method achieves state-of-the-art safety steering benchmarks and can also function as a structure-preserving image-editing technique, delivering high semantic fidelity while remaining model-agnostic and inversion-free.
arXiv:2609.39265v1 Announce Type: new Abstract: The Segment Anything Model (SAM) achieves remarkable performance in visual segmentation. The latest SAM3 extends promptable segmentation to concept-lev...
arXiv:2609.40317v1 Announce Type: new Abstract: While talking head generation has advanced rapidly, generating natural listener behavior in dyadic conversations, which know when to react, how to reac...
arXiv:2609.38397v1 Announce Type: new Abstract: Virtual clients offer a cost-effective approach to support applications such as A/B testing, recommender system development, and interface evaluation....
arXiv:2609.38411v1 Announce Type: new Abstract: Leading AI developers have reported agents acting beyond their approved limits, which a United Nations panel described as an early warning of loss of h...
arXiv:2609.39107v1 Announce Type: new Abstract: Large Language Models (LLMs) have been applied in various fields. However, ensuring compliance and safety of LLMs, such as avoiding discrimination and...