{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ATISBXMIVWYMBZDHB4K3TO3P5V","short_pith_number":"pith:ATISBXMI","schema_version":"1.0","canonical_sha256":"04d120dd88adb0c0e4670f15b9bb6fed5351db34cfae69b1aae5e9f9fca03034","source":{"kind":"arxiv","id":"2411.02571","version":2},"attestation_state":"computed","paper":{"title":"MM-Embed: Universal Multimodal Retrieval with Multimodal LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.IR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bryan Catanzaro, Chankyu Lee, Jimmy Lin, Mohammad Shoeybi, Sheng-Chieh Lin, Wei Ping","submitted_at":"2024-11-04T20:06:34Z","abstract_excerpt":"State-of-the-art retrieval models typically address a straightforward search scenario, in which retrieval tasks are fixed (e.g., finding a passage to answer a specific question) and only a single modality is supported for both queries and retrieved results. This paper introduces techniques for advancing information retrieval with multimodal large language models (MLLMs), enabling a broader search scenario, termed universal multimodal retrieval, where multiple modalities and diverse retrieval tasks are accommodated. To this end, we first study fine-tuning an MLLM as a bi-encoder retriever on 10"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2411.02571","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-04T20:06:34Z","cross_cats_sorted":["cs.AI","cs.CV","cs.IR","cs.LG"],"title_canon_sha256":"96771c02bf2a60e351e6fc8c20a67db6a582854cf1187ccadff2baa9537c3354","abstract_canon_sha256":"15e3a16a39a618cfcaf26490f29f49ba5fe7689708b44cf0f96caa0276617d6b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:18:10.882636Z","signature_b64":"mErjINKaJe8aRolLwcVZhle/1RFQApT+1mGZjW3QIXJzKAFL439oLX1c6EttskV6GoJM7hYCqkhjkp8YqcklCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"04d120dd88adb0c0e4670f15b9bb6fed5351db34cfae69b1aae5e9f9fca03034","last_reissued_at":"2026-07-05T10:18:10.882140Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:18:10.882140Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MM-Embed: Universal Multimodal Retrieval with Multimodal LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.IR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bryan Catanzaro, Chankyu Lee, Jimmy Lin, Mohammad Shoeybi, Sheng-Chieh Lin, Wei Ping","submitted_at":"2024-11-04T20:06:34Z","abstract_excerpt":"State-of-the-art retrieval models typically address a straightforward search scenario, in which retrieval tasks are fixed (e.g., finding a passage to answer a specific question) and only a single modality is supported for both queries and retrieved results. This paper introduces techniques for advancing information retrieval with multimodal large language models (MLLMs), enabling a broader search scenario, termed universal multimodal retrieval, where multiple modalities and diverse retrieval tasks are accommodated. To this end, we first study fine-tuning an MLLM as a bi-encoder retriever on 10"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.02571","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2411.02571/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2411.02571","created_at":"2026-07-05T10:18:10.882195+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.02571v2","created_at":"2026-07-05T10:18:10.882195+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.02571","created_at":"2026-07-05T10:18:10.882195+00:00"},{"alias_kind":"pith_short_12","alias_value":"ATISBXMIVWYM","created_at":"2026-07-05T10:18:10.882195+00:00"},{"alias_kind":"pith_short_16","alias_value":"ATISBXMIVWYMBZDH","created_at":"2026-07-05T10:18:10.882195+00:00"},{"alias_kind":"pith_short_8","alias_value":"ATISBXMI","created_at":"2026-07-05T10:18:10.882195+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":22,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24094","citing_title":"Universal Guideline-Driven Image Clustering via a Hybrid LLM Agent","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20280","citing_title":"ELVA: Exploring Ranking-Driven Universal Multimodal Retrieval","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13141","citing_title":"Rethinking RAG in Long Videos: What to Retrieve and How to Use It?","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12215","citing_title":"MLT-Dedup: Efficient Large-Scale Online Video Deduplication via Multi-Level Representations and Spatial-Temporal Matching","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28329","citing_title":"$M^3 QuestionIng$: Multi-modal Multi-span Medical Question Answering","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24530","citing_title":"Unveil: Unified Visual-Textual Integration and Distillation for Multi-modal Document Retrieval","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27295","citing_title":"Gemini Embedding 2: A Native Multimodal Embedding Model from Gemini","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30027","citing_title":"DocRetriever: A Plug-and-Play Framework for Multimodal Document Retrieval with Comprehensive Benchmark","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.29606","citing_title":"HiKEY: Hierarchical Multimodal Retrieval for Open-Domain Document Question Answering","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16638","citing_title":"TTE-Flash: Accelerating Reasoning-based Multimodal Representations via Think-Then-Embed Tokens","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18434","citing_title":"TIGER-FG: Text-Guided Implicit Fine-Grained Grounding for E-commerce Retrieval","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2507.04590","citing_title":"VLM2Vec-V2: Advancing Multimodal Embedding for Videos, Images, and Visual Documents","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2509.24621","citing_title":"FreeRet: MLLMs as Training-Free Retrievers","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2512.13511","citing_title":"Adapting MLLMs for Nuanced Video Retrieval","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13277","citing_title":"Utility-Oriented Visual Evidence Selection for Multimodal Retrieval-Augmented Generation","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02073","citing_title":"PLUME: Latent Reasoning Based Universal Multimodal Embedding","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13710","citing_title":"SLQ: Bridging Modalities via Shared Latent Queries for Retrieval with Frozen MLLMs","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23321","citing_title":"MMEB-V3: Measuring the Performance Gaps of Omni-Modality Embedding Models","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22280","citing_title":"Beyond Chain-of-Thought: Rewrite as a Universal Interface for Generative Multimodal Embeddings","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12148","citing_title":"ViLL-E: Video LLM Embeddings for Retrieval","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11095","citing_title":"Bottleneck Tokens for Unified Multimodal Retrieval","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13710","citing_title":"SLQ: Bridging Modalities via Shared Latent Queries for Retrieval with Frozen MLLMs","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V","json":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V.json","graph_json":"https://pith.science/api/pith-number/ATISBXMIVWYMBZDHB4K3TO3P5V/graph.json","events_json":"https://pith.science/api/pith-number/ATISBXMIVWYMBZDHB4K3TO3P5V/events.json","paper":"https://pith.science/paper/ATISBXMI"},"agent_actions":{"view_html":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V","download_json":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V.json","view_paper":"https://pith.science/paper/ATISBXMI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.02571&json=true","fetch_graph":"https://pith.science/api/pith-number/ATISBXMIVWYMBZDHB4K3TO3P5V/graph.json","fetch_events":"https://pith.science/api/pith-number/ATISBXMIVWYMBZDHB4K3TO3P5V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V/action/storage_attestation","attest_author":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V/action/author_attestation","sign_citation":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V/action/citation_signature","submit_replication":"https://pith.science/pith/ATISBXMIVWYMBZDHB4K3TO3P5V/action/replication_record"}},"created_at":"2026-07-05T10:18:10.882195+00:00","updated_at":"2026-07-05T10:18:10.882195+00:00"}