A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Christoph Strecha, Alex M. Bronstein, Michael M. Bronstein, et al.
LDAHash maps high-dimensional SIFT-like descriptors into compact binary strings via learned linear projections, enabling efficient Hamming-space matching with reduced storage.
Priyanka Banerjee, Andreas Philipp Eckert, Anna K. Schrey, et al.
ProTox-II is a freely available webserver that predicts 33 toxicity endpoints using molecular similarity, pharmacophores, fragment propensities, and machine-learning models.
Ruichen Zhang, Hongyang Du, Yinqiu Liu, et al.
Proposes an interactive AI framework with pluggable LLM and RAG modules for next-generation network management and optimization, demonstrating effective optimization problem design.
Boci Peng, Yun Zhu, Yongchao Liu, et al.
The first comprehensive survey of Graph Retrieval-Augmented Generation (GraphRAG), formalizing its workflow into graph-based indexing, graph-guided retrieval, and graph-enhanced generation.
Juan R. Terven, Diana Margarita Córdova Esparza, Julio-Alejandro Romero-González, et al.
A comprehensive survey of loss functions and metrics in deep learning, offering a unified framework, multi-loss insights, and guidance for modern tasks like RAG.
Jan Kocoń, Igor Cichecki, Oliwier Kaszyca, et al.
Automated evaluation of ChatGPT and GPT-4 on 25 NLP tasks reveals ~25% average quality loss versus SOTA, with harder tasks showing greater degradation.
Unknown
Neutex introduces neural texture mapping for volumetric neural rendering, enabling realistic image synthesis by leveraging volumetric neural rendering techniques.
Unknown
This paper benchmarks tool-using agents under tool-environment unreliability, revealing their fragility and advancing evaluation along three dimensions.
Unknown
This paper reviews synthetic data as a solution to supply chain machine learning challenges, covering methods, applications, and research opportunities.
Unknown
TechRAG proposes an evidence-gated multimodal agentic RAG architecture for technical literature reasoning, motivated by limitations of basic RAG.
Unknown
HAM-RAG formulates structure-faithful multimodal RAG and introduces HAM-Bench to evaluate hierarchy-aware interleaved generation.
Zhan-Lun Chang, Dong-Jun Han, Seyyedali Hosseinalipour, et al.
A two-stage multimodal RAG framework that separates multimodal query translation from generator-aligned reranking to bridge the semantic-utility gap.