The Archive

H3DNAS compresses 3D point cloud models directly on ONNX binaries for edge deployment without source code access, targeting NVIDIA Jetson Orin Nano.

Hybrid hallucination detection for black-box LLMs combines semantic entropy and token log-probability uncertainty to reduce false positives and missed fabrications.

Multi-objective RL framework balances traditional risk returns with ESG mandates across divergent rating methodologies for portfolio optimization.

oHC stabilizes Hyper-Connections via orthogonal SO(4) quaternion constraints on residual mixing matrices to bound layer-wise rescaling.

OpenAI and its CEO Sam Altman are facing 30 new lawsuits that accuse them of providing "substantial assistance and encouragement" to the suspect in Canada's Tumbler Ridge school shooting, as reported earlier by TechCrunch. The new wave of lawsuits was filed in a California federal court on Wednesday by the students, teachers, and the principal in the school at the time of the shooting. Similar to the lawsuits filed by Tumbler Ridge victims' families in April, these lawsuits claim OpenAI failed to take action after its automated review system flagged conversations that the alleged shooter, Jes...

Theoretical analysis of correlation gap bounds for set functions under pairwise independence, resolving conjecture limits for small dimensions.

New York City mayor Zohran Mamdani has announced a new policy today that will ban younger schoolchildren from using AI in classrooms. The one-year moratorium, effective in the 2026-2027 school year, will impact about 600,000 public school students in 2-K through eighth grade and is being introduced alongside additional limits on digital devices and a small pilot program for AI education tools. The new rules being introduced by NYC's Department of Education will bar teachers from using AI tools to grade assignments, and keep students from using the technology in the classroom until they enter ...

Multi-shell Leech-lattice decoder implementation for 2-bit LLM quantization with GPU kernels optimized for serving-phase GEMV at batch 1.

WinoQueer-NL benchmark measures anti-LGBTQ+ bias in Dutch language models via 42,906 culturally-validated stereotype sentences.

Loom aggregates conflicting textual hypotheses into consensus via embedding-space reweighting, deployed for industrial root cause analysis.

Differentiable optimization layer enables gradient-based planning for data center electricity costs by backpropagating through market-clearing constraints.

Anthropic publishes system prompts for Claude.ai and mobile apps, with version history showing evolving restrictions on song lyric reproduction.

TaRA improves LoRA initialization for efficient fine-tuning by incorporating training dynamics rather than static weight analysis.

Study quantifies prediction variance and complementarity in knowledge graph link prediction models, revealing model-specific knowledge capture gaps.

This is Adobe's second acquisition out of India after Rephrase.ai in 2023

LLMs automate vulnerability injection into Solidity smart contracts to generate ground-truth datasets for testing vulnerability detection tools across 49 OpenSCV types.

Pseudo-triplet construction pipeline enables direction-following text-to-speech generation without paired training data of script modifications and delivery directions.

EM-based approach estimates source distribution of simulator parameters by training amortized posteriors on fresh simulations rather than fixed surrogates.

As companies scale, the technology supporting operations can become a liability just as quickly as it becomes an asset. Disconnected systems, site-specific tools, spreadsheets, and manual workarounds can create data silos that make it harder to spot problems early, coordinate responses, and make decisions with confidence. For Jabil, a global manufacturing company with more than…

Theoretical analysis distinguishes novelty and diversity as components of creativity in human-AI collectives, challenging singular accounts of generative AI's cultural impact.

Multimodal speech-language framework predicts loneliness in older adults from telephone interviews using psycholinguistic features and topic modeling.

Replicates seminal economic experiments with LLM agents in double auction markets to test whether human-designed mechanisms deliver efficient resource allocation.

Learning-based attacks reconstruct encrypted point clouds from unencrypted coordinates, undermining selective coordinate encryption for volumetric video security.

Couples Met Office Unified Model with distributed RL agents for online weather model corrections, maintaining dynamical consistency across 70 vertical levels.

ProbeMatchDTI uses probe-driven pattern matching to improve drug-target interaction prediction by retaining weak binding-relevant signals in biochemical representations.

Answer-verified multi-teacher distillation identifies correct teacher per sample rather than per domain, improving multi-domain LLM integration from domain-matched supervision.

TrajMind uses chained LoRAs for fast urban trajectory anomaly diagnosis, balancing low-latency monitoring with semantic richness.

Controlled study of graph representations for GNN-based power grid control in L2RPN shows task granularity matters more than representational richness.

Proposes fine-grained anomaly detection dataset and difference-fusion framework for enhancement artifacts in social media images.