List your company

Knowledge-Based Visual Question Answering on E-VQA Single-Hop

59.9AccuracyPMSR
9.9822.9435.948.86Aug 31, 2025Oct 11, 2025Nov 21, 2025Jan 1, 2026Feb 11, 2026Mar 24, 2026May 5, 2026

Evaluation results

MethodLinks
PMSRRetriever=SigLIP2-g, Model=Gemini-2.5-Flash2025.0859.9
OMGMRetriever=EVA-CLIP-8B, Model=GPT-4o, Reranking=true2025.0851.2
OMGMRetriever=EVA-CLIP-8B, Model=LLaVA-1.5-7B, Reranking=true2025.0850.2
DBAgent (SFT: E-VQA)LLM=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B + bge-m32026.0446.5
PMSRRetriever=SigLIP2-g, Model=Qwen3-VL-8B2025.0846.4
DBAgent (SFT: Mixed)LLM=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B + bge-m32026.0446
DBAgent (SFT: InfoSeek)LLM=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B + bge-m32026.0445.1
PMSRRetriever=SigLIP2-g, Model=Qwen3-VL-4B2025.0840.9
CgRAGLLM/MLLM=InternVL3-8B, Retriever=EVA-CLIP-8B, Feature=V+T2026.0540.4
VLM-PRFLLM=InternVL3-8B, Retriever=EVA-CLIP-8B2026.0440.1
Wiki-PRF(w/ RL)Retriever=EVA-CLIP-8B, Model=VLM-PRF-7B2025.0840.1
VLM-PRF w/ RLLLM/MLLM=InternVL3-8B, Retriever=EVA-CLIP-8B, Feature=V+T2026.0540.1
CgRAGLLM/MLLM=LLaVA-NeXT-7B, Retriever=EVA-CLIP-8B, Feature=V+T2026.0539.9
CgRAGLLM/MLLM=LLaVA-NeXT-8B, Retriever=EVA-CLIP-8B, Feature=V+T2026.0539.9
CgRAGLLM/MLLM=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B, Feature=V+T2026.0539.8
MMKB-RAGRetriever=PreFLMR ViT-G, Model=Qwen2-VL-7B, Reranking=true2025.0839.7
MMKB-RAGLLM/MLLM=Qwen2-VL-7B, Retriever=EVA-CLIP-8B, Feature=V2026.0539.7
CgRAGLLM/MLLM=Qwen2-VL-7B, Retriever=EVA-CLIP-8B, Feature=V+T2026.0539.6
KU-RAGLLM/MLLM=GPT-4o, Retriever=EVA-CLIP-8B, Feature=V+T2026.0538.3
VLM-PRFLLM=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B2026.0437.1
ReflectiVALLM=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B2026.0436.8
EchoSightRetriever=EVA-CLIP-8B, Model=Mistral-7B, Reranking=true2025.0835.5
ReflectiVARetriever=EVA-CLIP-8B, Model=LLaVA-MORE-8B2025.0835.5
ReflectiVALLM/MLLM=LLaMA-3.1-8B, Retriever=EVA-CLIP-8B, Feature=V2026.0535.5
DPRLLM=Multi-passage BERT, Retriever=CLIP ViT-B/322026.0429.1
DPR_V+TLLM/MLLM=Multi-passage BERT, Retriever=CLIP ViT-B/32, Feature=V+T2026.0529.1
VLM-PRFLLM/MLLM=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B, Feature=V+T2026.0528.9
RET-2Retriever=RET-2, Model=LLaVA-MORE-8B2025.0828.5
ReflectVALLM=LLaMA-3.1-8B, Retriever=EVA-CLIP-8B2026.0428
GPT-4VLLM=GPT-4V, Retriever=–2026.0426.9
GPT-4VMode=Zero-shot MLLMs2026.0526.8
EchoSightLLM/MLLM=LLaMA-3.1-8B, Retriever=EVA-CLIP-8B, Feature=V2026.0526.4
DBAgent (w/o SFT)LLM=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B + bge-m32026.0423.6
LLaVA-NeXT-8BMode=Zero-shot MLLMs2026.0522.5
LLaVA-NeXT-7BMode=Zero-shot MLLMs2026.0522.1
Wiki-LLaVARetriever=CLIP-ViT-L, Model=LLaVA-1.5-7B2025.0821.8
Qwen2.5-VL-7B (Base)LLM=Qwen2.5-VL-7B (Base), Retriever=–2026.0421.7
Qwen2.5-VL-7BMode=Zero-shot MLLMs2026.0521.7
EchoSightLLM=Mistral-7B/LLaMA-3-8B, Retriever=EVA-CLIP-8B2026.0419.4
EchoSightLLM/MLLM=Mistral-7B/LLaMA-3-8B, Retriever=EVA-CLIP-8B, Feature=V2026.0519.4
Wiki-LLaVALLM/MLLM=LLaMA-3.1-8B, Retriever=CLIP ViT-L/14 + Contriever, Feature=T2026.0518.3
Qwen2.5-VL-3BMode=Zero-shot MLLMs2026.0517.9
Wiki-LLaVALLM=LLaVA-v1.5-7B, Retriever=CLIP ViT-L/14 + Contriever2026.0417.7
Wiki-LLaVALLM/MLLM=Vicuna-7B, Retriever=CLIP ViT-L/14 + Contriever, Feature=T2026.0517.7
LLaVA-v1.5LLM=Vicuna-7B, Retriever=–2026.0416.3
LLaVA-1.5LLM/MLLM=Vicuna-7B, Mode=Zero-shot MLLMs2026.0516.3
Qwen2-VL-7BMode=Zero-shot MLLMs2026.0516.2
LLaVA-1.5LLM/MLLM=LLaMA-3.1-8B, Mode=Zero-shot MLLMs2026.0516
BLIP-2LLM=Flan-T5XL, Retriever=–2026.0412.6
BLIP-2LLM/MLLM=Flan-T5XL, Mode=Zero-shot MLLMs2026.0512.6
InstructBLIPLLM=Flan-T5XL, Retriever=–2026.0411.9
InstructBLIPLLM/MLLM=Flan-T5XL, Mode=Zero-shot MLLMs2026.0511.9