4 papers
GOAT: A Training Framework for Goal-Oriented Agent with Tools
Hyunji Min, Sangwon Jung, Junyoung Sung +3
Current approaches rely on zero-shot evaluation due to the absence of training data; while proprietary models such as GPT-4 exhibit strong reasoning capabilities, smaller open-sour…
Breaking the Visual Shortcuts in Multimodal Knowledge-Based Visual Question Answering
Dosung Lee, Sangwon Jung, Boyoung Kim +4
Existing Multimodal Knowledge-Based Visual Question Answering (MKB-VQA) benchmarks suffer from "visual shortcuts", as the query image typically matches the primary subject entity o…
ReTAG: Retrieval-Enhanced, Topic-Augmented Graph-Based Global Sensemaking
Boyoung Kim, Dosung Lee, Sumin An +2
Recent advances in question answering have led to substantial progress in tasks such as multi-hop reasoning. However, global sensemaking-answering questions by synthesizing informa…
ReSCORE: Label-free Iterative Retriever Training for Multi-hop Question Answering with Relevance-Consistency Supervision
Dosung Lee, Wonjun Oh, Boyoung Kim +3
Multi-hop question answering (MHQA) involves reasoning across multiple documents to answer complex questions. Dense retrievers typically outperform sparse methods like BM25 by leve…