New York passed the NY FAIR News Act, requiring news organizations operating in the state to provide clear disclosures for content substantially or wholly generated by AI and to implement safeguards protecting journalist sources and confidential materials from AI access. The bill is heading to Gov. Kathy Hochul for signature.
Illinois laws taking effect July 1 include House Bill 3851, which treats posting unauthorized AI-generated images as cyberbullying in schools. The article also notes other enacted laws with different subject areas.
A former New York City Council candidate was charged with forgery for allegedly using AI-generated images and posts to fabricate political endorsements and news stories on social media. The charges relate to third-degree forgery and possession of forged instruments under New York law.
UN chief António Guterres urged AI companies to disclose their environmental impact and launched an AI Environmental Transparency Initiative, including a call to power data centres with renewable energy by 2030.
UNESCO-led ministerial summit in the Dominican Republic supported ethical AI governance and regional cooperation aligned with the UNESCO Recommendation on the Ethics of Artificial Intelligence. Dominican authorities also announced a forthcoming national AI Code of Ethics.
Attention is a core operation in large language models (LLMs). We present BD Attention (BDA), a lossless algorithmic reformulation of attention. BDA is enabled by a simple matrix identity from Basis Decomposition (BD), which restructures multi-head projections into a compact form while preserving exact outputs. Unlike I/O-aware system optimizations such as FlashAttention, BDA provides a mathematic
Safety alignment in Large Language Models is critical for healthcare; however, reliance on binary refusal boundaries often results in over-refusal of benign queries or unsafe compliance with harmful ones. While existing benchmarks measure these extremes, they fail to evaluate Safe Completion: the model's ability to maximise helpfulness on dual-use or borderline queries by providing safe, high-leve
Self-supervised learning (SSL) and diffusion models have respectively advanced representation learning and generative modeling for high-dimensional 3D visual data, yet they are often developed as separate paradigms. Their unification remains challenging under multi-source heterogeneity, as anatomical content must be preserved for analysis while acquisition-related style varies across centers and a
Non-rigid 3D shape matching is a fundamental task in computer vision and graphics. In this paper, we propose a hybrid self-supervised method based on a coarse-to-fine strategy, which ensures consistency between the coarse mapping and the refined correspondence produced by our refinement module. The architecture features a dual-branch design, consisting of two symmetric functional map learning stre
To reduce memory consumption during LLM inference, a handful of methods have been proposed for KV cache pruning. While these techniques can accomplish lossless memory reduction on many datasets, they often hinge on an under-emphasized condition: an input/domain-specific threshold for KV cache budget needs to be pre-determined to achieve the optimal performance. However, such input-sensitive design
Capital is concentrating at the top end, with a large private round and a major public listing dominating the tape. The mix points to continued investor appetite for infrastructure and frontier AI exposure rather than a broad seed-stage spread.
Capital is concentrating at the top end, with a large private round and a major public listing dominating the tape. The mix points to continued investor appetite for infrastructure and frontier AI exposure rather than a broad seed-stage spread.
Capital is concentrating at the top end, with a large private round and a major public listing dominating the tape. The mix points to continued investor appetite for infrastructure and frontier AI exposure rather than a broad seed-stage spread.
Named moves are still sparse, but the hires and founding activity point to teams being assembled around frontier research and new company formation. The signal today is less churn than deliberate staffing around emerging AI bets.
Named moves are still sparse, but the hires and founding activity point to teams being assembled around frontier research and new company formation. The signal today is less churn than deliberate staffing around emerging AI bets.
Named moves are still sparse, but the hires and founding activity point to teams being assembled around frontier research and new company formation. The signal today is less churn than deliberate staffing around emerging AI bets.
Benchmark activity is narrow but useful: the cycle adds fresh scores on reasoning and math-style tests without a broader leaderboard shake-up. That keeps attention on incremental capability tracking rather than a new top-line reset.
Benchmark activity is narrow but useful: the cycle adds fresh scores on reasoning and math-style tests without a broader leaderboard shake-up. That keeps attention on incremental capability tracking rather than a new top-line reset.
Benchmark activity is narrow but useful: the cycle adds fresh scores on reasoning and math-style tests without a broader leaderboard shake-up. That keeps attention on incremental capability tracking rather than a new top-line reset.