Uses paired image-domain retrieval and text augmentation to improve factuality of 3D brain MRI radiology report generation.
Maps homogeneity bias across temperature and top-p parameters in seven open-weight instruction-tuned LLMs (7-20B), finding bias robust to decoding settings.
RL-based optimal control framework balancing safety constraints and performance in learning-based control for nonlinear systems.
Addresses long-tailed distribution in class-conditional diffusion models to improve image diversity for underrepresented classes.
SyncLoop dual-loop framework for self-improving multimodal LLMs with iterative refinement for mathematical reasoning without manual dataset curation.
CNN surrogate models with domain decomposition for predicting blood flow in stenosed arteries using Schwarz method.
Proposes BOFA method for class-incremental learning using CLIP vision-language models with orthogonal low-rank fusion to prevent forgetting.
Develops formal convergence theorem for LLM-verifier systems integrating formal verification tools with LLMs for reliable software verification.
Uses transformers for gravitational-wave parameter estimation from noisy detector signals with flexible handling of data variations.
Theoretical analysis of learning parameters for symmetric stochastic linear dynamical systems from limited observations.
Graph neural network for jet classification in particle physics using kinematic variables. Improves interpretability in deep learning models.
Cross-platform analysis comparing climate discourse on Meta paid ads versus public social media feeds using theme-based interpretable methods.
Hybrid temporal graph network and SEAL model for link prediction in dynamic sparse networks capturing temporal dependencies.
Streamlined spectral algorithm for community detection in stochastic block models achieving information-theoretic bounds with fewer computational steps.
Public benchmark for tandem mass spectrum prediction with standardized metadata and preprocessing to enable fair deep learning model comparison.
Multifidelity Gaussian process approach for surrogate modeling from scarce data using low and high-fidelity training samples.
Theoretical analysis of grokking phenomenon in feature learning kernels, studying when data symmetry breaking enables generalization on algebraic tasks.
Lossless KV cache compression technique for disaggregated LLM serving that reduces transfer bottleneck between prefill and decode workers.
Systematic evaluation of six defenses against persistent memory attacks on stateful LLM agents across architectural layers with nine open-source models.
Zero-shot cross-lingual confidence estimation method for multilingual LLMs, demonstrating language-transferable confidence features without retraining.
Real-time human activity recognition system for smart homes using trajectory-guided localization to handle continuous streaming sensor data.
Theoretical analysis of Classification and Regression Trees through Bregman divergences, unifying impurity measures across different statistical models.
Skill-MAS framework for automatic multi-agent system generation from LLMs that evolves meta-skills to balance model capability with experience retention.
Diffusion-based approach for sequence labeling in NLP that extends linear-chain CRFs with improved expressivity through approximate structured inference.
Framework and evaluation of failure modes in AI coding agents, addressing underspecification, capability errors, and mitigation strategies.
Energy-aware optimization framework for on-device LLM inference that reduces power consumption by adjusting processor frequencies while maintaining performance.
Framework for accelerating video diffusion model inference through instance-specific optimization strategies across different hardware configurations.
Research on model selection for probabilistic forecasting using proper scoring rules on time series datasets, focusing on aggregation methods.
Analysis of security tool adoption in AI coding agents. Examines how LLM-based agents handle security features in SDLC.
Opinion piece on future of keyboardless computing and natural interfaces. Speculative, no technical content.
Video presentation on LXM pseudorandom number generators with improved statistical properties. Research topic.
Personal essay on using Claude Code and AI tools for writing and learning. Reflective opinion piece without technical depth.
LodeDB: embedded vector database optimized for local RAG. GPU-accelerated, sub-millisecond latency, compatible with LangChain and LlamaIndex.
Discussion on how AI is changing software development profession and coding practices. Observational thread.
GitHub issue about home feed not displaying starred repository releases. Bug report.
chrome-use: Open-source tool enabling AI agents to control real Chrome browsers with existing logins and anti-detection. Part of *-use family.
SCBKR: local LLM control framework with owner-signed responsibility chains. Governance layer for local AI ensuring user verification and audit trails.
Open-source personality-based dating app with conversation-first approach. AGPL-3.0 licensed full-stack application.
TronBrowser: Open-source privacy-first AI-native web browser on Ungoogled Chromium with built-in AI sidebar and agent CLI. MIT licensed.
Economic research on AI agents for knowledge work. Agents perform long-horizon delegated tasks versus short chatbot interactions.
Technical case study: AI desktop agent failure due to driver incompatibility creating circular dependency. Real-world deployment issue analysis.
Promptctl: Git-like version control tool for LLM prompts. Tracks, diffs, and rolls back prompts with CLI. Open-source in Go.
Experience report on token costs and code style when using Claude for feature development. Discusses optimization patterns.
News headline: Google Gemini 3.5 Pro release delayed to July. No additional details.
Helper script enabling use of Claude's Chrome extension from Cursor IDE for browser tab analysis and debugging.
S3-based durable filesystem layer for AI agents. Syncs memory files across platforms. Rust implementation with Python/TypeScript SDKs and CLI.
Analysis of AI prototyping tools and multiplayer collaboration challenges. Examines export compatibility and agent coordination issues in editors.
Technical guide on monitoring and tracing LLM agents. Covers telemetry for multi-turn, multi-tool agent execution and error detection.
Analysis of developer morale and identity issues amid AI coding tool adoption. Reports divide between 'lazy' and 'thoughtful' engineers.
Social media platform announcement with no AI/tech relevance. Text-only network focusing on chronological feeds.