Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2609.21597v1 Announce Type: new Abstract: Monocular 6-DoF pose estimation of non-cooperative targets is important for on-orbit servicing and debris removal. A single-image estimator can confuse...
arXiv:2609.21938v1 Announce Type: new Abstract: Transformer-based models have recently achieved strong performance on 3D reconstruction from images, and recent works extend them to process video stre...
arXiv:2609.20892v1 Announce Type: cross Abstract: Closed-loop visual servoing requires predictions that indicate whether an action reduces task error, not only whether the action is plausible. We cal...
arXiv:2406.08775v2 Announce Type: replace Abstract: Labels are the cornerstone of supervised machine learning algorithms. Most visual recognition methods are fully supervised, using bounding boxes or...
arXiv:2609.20844v1 Announce Type: new Abstract: Deepresearch (DR) agents interact with real-world web environments through multi-turn search and visit, causing their contexts to grow rapidly over tim...
arXiv:2609.21673v1 Announce Type: new Abstract: Probabilistic Graphical Models (PGMs), especially Bayesian Networks (BNs), expose directed structure and probabilistic parameters, making them natural...
arXiv:2609.21470v1 Announce Type: new Abstract: Sparse representation formulates the environment perception for the end-to-end driving system as a set of discrete elements like objects and lane lines...
The paper presents a framework and results of a multi-dimensional analysis of GPTKB v1.5, a 100‑million‑fact knowledge base elicited from GPT‑4.1. It shows that the LLM’s factual knowledge differs markedly from established knowledge bases and that its accuracy is lower than suggested by prior benchmarks. The study also identifies inconsistency, ambiguity, and hallucinations as major issues, pointing to future research directions in neuro‑symbolic AI for extracting, consolidating, and verifying factual LLM knowledge.
arXiv:2607. 00876v3 Announce Type: replace-cross Abstract: Private continual counting is a fundamental problem in differential privacy: given a binary stream of length $n$, where each $1$ corresponds to the contribution of one individual, the goal is to release all running counts while protecting the privacy of each individual.
Previous work on Egyptian Arabic in NLP has focused largely on the prestigious Cairene Egyptian Arabic (CEA) dialect, resulting in a lack of representation for the less prestigious Sa'idi Egyptian Ara...
LLM explainers are increasingly attached to autonomous agents as runtime oversight, with operators reading a generated account of the agent's beliefs and actions rather than its internal state. We aud...
We present M3GA-Wild, the first benchmark for multi-modal, multi-session ground-to-aerial place recognition in forests. M3GA-Wild unifies and extends existing forest localisation datasets, providing a...
Accurate and timely detection of degradation in lithium-ion batteries is crucial to ensure safety, reliability, and longevity in high-demand applications such as electric vehicles and energy storage s...
OpenAI has announced the Australian Youth Safety Blueprint, a six-pillar roadmap designed to create safer AI experiences for young people. The initiative aims to protect and empower youth by outlining specific safety measures and guidelines. It represents a focused effort to address the unique needs and risks faced by younger users of AI technology.
LaSR (Latent Speech Reasoning) is a new training paradigm for context‑aware speech recognition that uses a latent reasoning trajectory instead of explicit intermediate tokens. It aligns chain‑of‑thought supervision around the acoustic region of target words and introduces latent reasoning periods for grounding context and guiding transcription transitions. Experiments on Fun‑Audio‑Chat show that LaSR improves terminology recognition without added latency and outperforms standard fine‑tuning baselines, demonstrating the promise of latent reasoning for efficient, context‑aware speech assistants.
The paper introduces Decoupled Contrastive Decoding (DCD), a method that separates drafting from verification in contrastive decoding. DCD uses an expert‑aligned lightweight proposer for drafting and applies contrastive correction only during verification, preserving the vanilla contrastive decoding distribution. Experiments on 8B models show that DCD achieves 1.65–1.95× speedups in greedy decoding and reduces MMLU proposal‑path latency by 5–12× compared to amateur‑coupled approaches.
UniPolicy is a unified objective‑specific policy framework for search advertising that jointly optimizes relevance, click propensity, and commercial value. It uses objective‑aware prefix tokens, sparse MoE‑LoRA routing, and residual FFNs to decouple parameters within a shared backbone, and constructs pairwise preferences from multi‑stage behavioral feedback to strengthen clicked candidates. In large‑scale offline tests and a 7‑day online A/B test, UniPolicy improves CTR by 0.71%, RPS by 1.58%, and advertising revenue by 1.32% while keeping serving latency stable.
The paper "Benchmarking LLM Compliance with China AI Generated Content Regulations" presents a new framework for evaluating large language models (LLMs) against China’s AI‑generated content compliance rules. Using 2,303 questions across six dimensions—including 203 self‑constructed constitutional queries—the authors assess 20 prominent LLMs, finding that international models show high compliance, especially on ideologically aligned dimensions. The study establishes a regulatory benchmark that allows the global AI community to test both Chinese and non‑Chinese LLMs under unified, legally grounded compliance criteria.
HerHealthEval is a controlled evaluation framework that tests multilingual understanding of women's-health communication across English, French, and Modern Standard Arabic. It provides six communicative forms—canonical, clinical, layperson, indirect or hedged, emotionally concerned, and deliberately under-specified—each conveying the same clinical concern except the under-specified form omits details to assess clarification needs. The study evaluates multilingual instruction models and QLoRA-adapted variants on tasks such as concern classification, risk calibration, clarification behavior, parse compliance, and cross-form consistency, revealing that high aggregate accuracy can mask safety-relevant failures and that language-invariant risk labels improve performance.
The paper introduces CL-DAF, a Cross‑Lingual Disease‑Alignment Framework that identifies acoustic features whose disease effects are consistent across languages. Using data from 201 English and 75 Bangla speakers, the authors construct a 272‑dimensional acoustic representation and find that 133 features reverse disease direction between languages, leading to poor cross‑lingual transfer (AUC 0.49). By isolating 26 disease‑aligned features, CL‑DAF improves cross‑lingual classification, achieving AUCs of 0.825 (English to Bangla) and 0.722 (Bangla to English).