A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Marcus Olivecrona, Thomas Blaschke, Ola Engkvist, et al.
A deep reinforcement learning method tunes a sequence-based generative model to design molecules with desired properties, achieving over 95% predicted activity for dopamine receptor type 2.
Reza Olfati‐Saber, Richard M. Murray
Establishes convergence analysis for consensus protocols in directed networks with fixed/switching topologies and time-delays, linking algebraic connectivity to performance.
Unknown
Introduces PRM BiasBench, a benchmark suite for systematically evaluating the robustness and failure modes of process reward models.
Karl Hanna, Chen Feng
This paper investigates whether label-free multiple-choice strategies can remove option-order bias in LLMs, finding they do not reliably improve accuracy and that withholding options is the bottleneck.
Ananya Sahu, Mohit Bansal, Elias Stengel-Eskin
CreativeInstruct teaches LLMs to balance creativity and quality via special [StartCreativity] spans, matching or exceeding diversity baselines without sacrificing quality.
Unknown
This paper argues that effective real-world LLM agents require optimizing productivity, proactivity, and personalization, and proposes a training approach to achieve this.
Nossa Iyamu
A deterministic, zero-model pipeline compiles passively captured screen activity into typed activity frames, reducing raw capture 86x and enabling 98.4% question-answering accuracy, while also measuring agent cost parameters.
Unknown
This paper reviews federated learning applications, highlighting its role as a decentralized privacy-preserving technology to address data silos and data sensitivity challenges.
Unknown
Introduces Vitabench, a benchmark for evaluating LLM agents on versatile interactive tasks in real-world applications, capturing complexity of handling extensive information.
Hongming Chen, Ola Engkvist, Yinhai Wang, et al.
This paper reviews the first wave of deep learning applications in drug discovery, covering bioactivity prediction, de novo molecular design, synthesis prediction, and biological image analysis.
Ioannis Maniadis Metaxas, Adrian Bulat, Alberto Baldrati, et al.
UltraViT is a latency-optimized vision encoder for LVLMs using a pyramidal architecture with heterogeneous spatial mixers and a two-stage generative pre-training strategy, achieving 1.7x speedup on-device.
Manuel Cossio
This paper provides a comprehensive taxonomy of LLM hallucinations, arguing their theoretical inevitability and emphasizing the need for robust detection, mitigation, and human oversight.