Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2608.23100v1 Announce Type: cross Abstract: Robot co-design via bi-level optimization couples within-lifetime controller learning for fitness evaluation with cross-generational morphological ev...
The paper introduces MemeMind, a large-scale dataset for detecting harmful memes that includes a detailed taxonomy and Chain-of-Thought reasoning annotations. It also proposes MemeGuard, a multimodal framework that uses a three-stage training strategy to improve visual understanding, reasoning, and discrimination of harmful content. Experiments show MemeGuard surpasses current state-of-the-art methods on MemeMind, advancing detection accuracy and interpretability.
arXiv:2506.10630v4 Announce Type: replace-cross Abstract: To advance time series forecasting (TSF), various methods have been proposed to improve prediction accuracy, evolving from statistical techni...
arXiv:2608.23458v1 Announce Type: cross Abstract: Interpretability is critical for machine learning models deployed in scientific space missions such as ESA's Ariel, where ground truth is unavailable...
Omni‑SafetyBench is a new benchmark designed to evaluate the safety of Omni‑Modal Large Language Models (OLLMs) that process visual, auditory, and textual data. It contains 23,328 test instances across 24 modality variations derived from 972 seed samples, and introduces metrics such as Safety‑score (based on Conditional Attack Success Rate and Conditional Refusal Rate) and Cross‑Modal Safety Consistency score. Evaluation of 11 state‑of‑the‑art OLLMs shows severe vulnerabilities, with only three models achieving a Safety‑score above 0.6 and safety degrading sharply for audio‑visual inputs, underscoring the need for improved safety alignment methods.
arXiv:2608.22950v1 Announce Type: new Abstract: Floating waste in inland waterways threatens aquatic ecosystems and requires timely monitoring under cluttered, multi-object conditions. Existing aquat...
arXiv:2608.22272v1 Announce Type: cross Abstract: Generative adversarial networks (GANs) can provide efficient image generation, while diffusion models offer high-quality image restoration but requir...
arXiv:2608.21444v1 Announce Type: new Abstract: Multi-drone systems are increasingly positioned for safety-critical missions such as search and rescue (SAR) and critical infrastructure monitoring. Ye...
arXiv:2608.03446v2 Announce Type: replace Abstract: Multilingual large language models (LLMs) have been shown to perform better on non-English classification tasks when the representations of the giv...
The article discusses the evolution of AI across three generations—from explicit logic to neural networks to large language models (LLMs)—and how LLMs introduce new systemic risks. It applies a forensic‑psychology profiling method to identify ten key features of LLMs, such as hallucinations, bias, and cognitive atrophy, revealing an entity that confabulates, amplifies user biases, and erodes human competence. The report concludes with a four‑pillar framework for AI resilience, emphasizing cognitive sovereignty, measurable control, partial autonomy, and openness to safeguard society.
The study examines how multilingual large language models (LLMs) produce outputs that differ across sociocultural contexts, highlighting that identity labels and source-language cues can mislead assessments of cultural grounding. Using a human‑validated, multi‑agent audit on 89,253 outputs from 12 LLMs in English, French, and Chinese across 18 occupations and three task conditions, the authors find that bias representation varies systematically by language and task. Removing direct identity cues reduces identity‑label prediction in English and Chinese but not in French, and the source language’s cultural context consistently receives the highest relevance score, though this signal weakens after translation or name masking. "whyItMatters":"The findings show that surface cues can obscure true cross‑cultural patterns, underscoring the need for careful audit designs to avoid misleading conclusions about bias in multilingual LLMs."
arXiv:2608.22762v1 Announce Type: new Abstract: Knowledge graph question answering (KGQA) is a key task for evaluating KG-augmented Large Language Models (LLMs), and complex KGQA that requires multi-...
arXiv:2608.22797v1 Announce Type: new Abstract: Background This study was designed to evaluate whether a domain-specific large language model (LLM) trained exclusively on patient education resources...
TEE-X is a TEE‑aware acceleration framework designed to run large vision models, such as Vision Transformers, entirely within Trusted Execution Environments. It introduces a sensitivity‑aware modularization technique and vectorization to overcome memory constraints and latency challenges on edge devices. The framework is validated on OP‑TEE for Arm TrustZone and optimized for the NVIDIA Jetson AGX Xavier, achieving GPU‑level inference latency with minimal accuracy‑latency trade‑offs.
arXiv:2608.22972v1 Announce Type: new Abstract: Surgical action triplet recognition constitutes a critical task in context-aware robot-assisted surgery, facilitating automatic surgical action percept...
arXiv:2312.10376v2 Announce Type: replace Abstract: Typical methods for visual prompt tuning follow the sequential modeling paradigm originating from NLP, learning a sequence of unordered parameteriz...
VISTA is a gradient‑based test‑time alignment framework designed for next‑scale visual autoregressive (VAR) image generation. It optimizes intermediate representations within the frozen transformer to enforce compositional constraints, without altering model weights or requiring extra training. Experiments on two benchmarks and two model scales show that VISTA improves compositional accuracy by up to 20% on a 2B backbone and 6% on an 8B backbone, while preserving image quality and enabling a smaller model to outperform a larger one.
arXiv:2608.22367v1 Announce Type: new Abstract: Diffusion multimodal large language models (dMLLMs) frequently produce long-form outputs marred by semantic drift and repetition, with quality generall...
The paper introduces Scaling Electronic Health Record Foundation Models for Population Health Management, a large‑scale model trained on billions of medical events from over 5 million patients in Taiwan and the United States. By aligning ICD codes across different health systems, the model achieves strong scaling and generalization across 11 chronic disease prediction tasks, outperforming tree‑based, general, and biomedical language models with high sensitivity at 99% specificity. It also demonstrates superior few‑shot performance on the EHRShot benchmark and shows that cross‑system alignment provides a stronger pretraining signal than single‑site duplication in data‑limited scenarios.
The paper evaluates using large language model (LLM) juries to review code generated from natural language queries, focusing on MySQL text-to-SQL tasks. It benchmarks 15 open models, selects the top six, and constructs unanimous committees of varying sizes to accept a query only when all members agree. The study finds that single-model judges are inconsistent, while small unanimous committees of strong models can reduce false accepts without discarding many correct queries, and that committee composition significantly influences performance.