Open Domain
Open-domain research focuses on developing AI systems capable of handling diverse, unstructured inputs and tasks without requiring extensive pre-training or fine-tuning for each specific domain. Current research emphasizes retrieval-augmented generation (RAG) methods, often incorporating knowledge graphs and vector stores to improve accuracy and reduce hallucinations, alongside advancements in masked diffusion transformers for efficient sound and image generation. This work is significant because it aims to create more adaptable and robust AI systems applicable across various fields, from e-commerce chatbots to autonomous driving and biomedical named entity recognition, ultimately improving the accessibility and effectiveness of AI technologies.
Papers
Public Domain 12M: A Highly Aesthetic Image-Text Dataset with Novel Governance Mechanisms
Jordan Meyer, Nick Padgett, Cullen Miller, Laura Exline
EvoCodeBench: An Evolving Code Generation Benchmark with Domain-Specific Evaluations
Jia Li, Ge Li, Xuanming Zhang, Yunfei Zhao, Yihong Dong, Zhi Jin, Binhua Li, Fei Huang, Yongbin Li
Domain-Specific Retrieval-Augmented Generation Using Vector Stores, Knowledge Graphs, and Tensor Factorization
Ryan C. Barron, Ves Grantcharov, Selma Wanna, Maksim E. Eren, Manish Bhattarai, Nicholas Solovyev, George Tompkins, Charles Nicholas, Kim Ø. Rasmussen, Cynthia Matuszek, Boian S. Alexandrov
MDSGen: Fast and Efficient Masked Diffusion Temporal-Aware Transformers for Open-Domain Sound Generation
Trung X. Pham, Tri Ton, Chang D. Yoo