Newest

Palomar is a new registry for Lean verified mathematics proofs, designed to help validate formal proofs using mechanical checks and AI-assisted methods.

Cerebras launches the CS-4, a rack-scale AI system with WSE-3 Turbo technology claiming up to 30x faster inference than GPUs, featuring a modular design for efficient hyperscale deployment.

OpenLogi is an open-source, local-first tool written in Rust for configuring Logitech mice over HID++, offering button remapping, DPI control, and per-app profiles without requiring an account or telemetry.

Meta is involved in a major trial that draws parallels to historical big tobacco cases, indicating significant legal scrutiny for the tech industry.

The paper introduces multi-byte prediction to speed up inference in byte-level language models by generating multiple bytes in parallel with minimal performance impact.

This paper proposes Agentic ESOpt, a method using evolution strategies to enable scalable full-parameter fine-tuning of long-horizon LLM agents with minimal GPU memory requirements.

The paper proposes RUPA, a framework that models LLM agent execution as a dependency graph to propagate uncertainty, improving failure detection and confidence estimation in long trajectories.

The paper introduces aDSL, a co-designed domain-specific language and multi-agent system that improve LLM-driven 3D program synthesis through relational operators and iterative feedback, enhancing robustness and controllability in 3D content creation.

StartupBench introduces a benchmark for evaluating general-purpose AI agents on real-world startup workflows, revealing that top models complete only about 30% of tasks due to gaps in complex instruction following and domain-specific expertise.

This paper presents a systematic scaling law study for text-to-image diffusion models, showing they scale predictably but require significantly more data per parameter than language models for optimal training.

ASI-Bench is a new benchmark designed to evaluate AI systems' capabilities in innovative exploration and autonomous scientific execution across 11 scientific domains, revealing current AI's heavy dependence on human guidance.

This paper evaluates indirect prompt injection risks in DeepSeek Harness using AI-Infra-Guard for controlled testing, finding notable attack success rates and recommending security controls.

GS-Voxel introduces a fitting-free framework to convert 3D Gaussian Splatting reconstructions into structured latents, enabling scalable generation of large-scale aerial 3D scenes via flow models and tiled inference.

The paper introduces a capability-driven data infrastructure with curriculum scheduling to train generalist image generation models using heterogeneous supervision for diverse generative tasks.

Energy-Guided Flow Matching improves generative image quality by using a moving endpoint and adaptive scheduling, achieving state-of-the-art FID scores with reduced training cost.

TAMP-Nav is a unified framework that enhances embodied navigation by aligning vision-language models with 2D visual prompting, selective reasoning with compressed memory, and policy optimization, achieving state-of-the-art performance with high runtime and sample efficiency.

FreeToken is an edge-native serving system that dynamically maps computation and model state onto heterogeneous local hardware to run large open-weight models on personal machines, enabling efficient execution of models up to 753B on a single GPU.

The paper introduces a common layer equation that unifies graph neural networks into seven components, enabling architectural comparison, theoretical analysis, and insights into issues like oversmoothing and expressivity.

MIT CSAIL researchers discovered 'attribution decay' in large generative AI models, where generated images often cannot be traced back to individual training data, using a novel diffusion ensemble architecture.

The article contrasts Vim's focus on user control through programmatic customization with VSCode's emphasis on consumption via heavier plugin development.