Last week coined the term “Black October“ after an onslaught of critical vulnerability disclosures, including PHP2Shell and GitLab, signified the moment when the cyber risks of AI models started to materialize.
The Data Security Council of India (DSCI) has published an India-focused guide on artificial intelligence (AI) liability for agentic AI systems. This guide clarifies how responsibility is allocated across the AI value chain if an agent operates autonomously or deviates from its tasks.
Researchers reveal that Chinese AI model GLM-5.3 has successfully executed complex cyberattacks in controlled tests, highlighting the potential dangers of open-weight AI systems and cybersecurity risks.
MIT's new LLM Election Observatory monitors in real time how major AI chatbots respond to questions about the 2026 midterm elections, revealing differences in framing, sources, and tone across user identities.
Cursor removed a compromised Claude Code extension from its marketplace on October 29. This came after researchers found out that the malicious extensions installed an SSH backdoor. An estimated 5,000 users may have been affected.
Cathie Wood just pulled millions out of two of this year's hottest AI chip winners and poured it into a battered broker and a pair of air taxi startups burning through cash. The trades raise a pointed question: does ARK see something the market is punishing,
Data attribution methods for Language Models (LMs) are typically evaluated using downstream tasks like fact retrieval. However, existing evaluations vary multiple dataset properties simultaneously, making it impossible to isolate exactly why an attribution algorithm fails. To diagnose these failures, we explore the design space of LLM data attribution, with a primary focus on evaluation. We introduce the Controlled Attribution Benchmark (CATT) for fact tracing, in which we fine-tune LLMs on entirely fictitious knowledge, providing unambiguous ground truth for attribution. This synthetic setup allows us to systematically alter one dataset configuration at a time, such as document redundancy and lexical distractors. With this configurable evaluation, we demonstrate that redundancy and distractors consistently degrade attribution performance. Rather than proposing a new estimator, we examine two design choices that are underexplored: the loss used to compute gradients and how per-token attribution scores are aggregated into a document score. We find that the true causal evidence within a training document is highly localized, but standard methods dilute this evidence into noise by…
The prompt is built automatically as "" * num_pages + "Multi page parsing.", and each page is encoded as one global view (base mode). The output contains one per input page, in the same order as image_files.
Text summarization is an important task for advancing natural language processing (NLP) research and applications in Indic languages. Despite progress in extractive and abstractive summarization, both monolingual summarization (MS) and cross-lingual summarization (CLS) remain…
Data augmentation improves visual recognition by exposing models to synthetic training examples that encourage generalization beyond the observed data. Mixed-sample methods extend this idea by combining multiple images. However, many common approaches either mix only pairs of…
Your team of AI agents, on your Mac. Agents that talk in channels, write code, review each other and use your tools, on the Claude Code, Codex, OpenCode or Pi you already have. Anything that touches the real world waits for your yes.
Topology-agnostic motion generation aims to synthesize motions for diverse characters with arbitrary skeletons. Unlike conventional motion generation, which assumes a fixed skeletal template, this setting must handle diverse joint configurations and inconsistent joint semantics…