{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:ENR67QO2PL3NE43ESIDJWBQHOF","short_pith_number":"pith:ENR67QO2","canonical_record":{"source":{"id":"2607.26947","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-07-29T14:15:55Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bf18220f5339ccbcd4236ed53c48f06c9bbed0db936970692051befeb40dc355","abstract_canon_sha256":"7a5ed24acf989c9bfa0699d2bd4dd79d7ac028d7ee4c339aa4cabe64c170adb3"},"schema_version":"1.0"},"canonical_sha256":"2363efc1da7af6d2736492069b060771679076928fcead58df7410722d8fc183","source":{"kind":"arxiv","id":"2607.26947","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.26947","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"arxiv_version","alias_value":"2607.26947v1","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.26947","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"pith_short_12","alias_value":"ENR67QO2PL3N","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"pith_short_16","alias_value":"ENR67QO2PL3NE43E","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"pith_short_8","alias_value":"ENR67QO2","created_at":"2026-07-30T01:23:08Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:ENR67QO2PL3NE43ESIDJWBQHOF","target":"record","payload":{"canonical_record":{"source":{"id":"2607.26947","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-07-29T14:15:55Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bf18220f5339ccbcd4236ed53c48f06c9bbed0db936970692051befeb40dc355","abstract_canon_sha256":"7a5ed24acf989c9bfa0699d2bd4dd79d7ac028d7ee4c339aa4cabe64c170adb3"},"schema_version":"1.0"},"canonical_sha256":"2363efc1da7af6d2736492069b060771679076928fcead58df7410722d8fc183","receipt":{"kind":"pith_receipt","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2363efc1da7af6d2736492069b060771679076928fcead58df7410722d8fc183","last_reissued_at":"2026-07-30T01:23:08.181432Z","signature_status":"unsigned_v0","first_computed_at":"2026-07-30T01:23:08.181432Z"},"source_kind":"arxiv","source_id":"2607.26947","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-30T01:23:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lcUK8m9oO6x74Rb+W6eVKLFzEmrxgZfxbQjAMlrG/pMJeRd1TtvdEIfV/K45FWSJvAr2Rt03spbZqJ2oPovwAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T13:34:10.642065Z"},"content_sha256":"4ffee3f3e02cd20034a3b766f55738e0b195876df3f09ae201352da7189292ca","schema_version":"1.0","event_id":"sha256:4ffee3f3e02cd20034a3b766f55738e0b195876df3f09ae201352da7189292ca"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:ENR67QO2PL3NE43ESIDJWBQHOF","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Progressive Multimodal Alignment for Continual Instruction Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Duzhen Zhang, Jiahua Dong, Qiaoyi Su, Tielin Zhang, Yahan Yu","submitted_at":"2026-07-29T14:15:55Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) rely on a projector to align visual representations with the language embedding space, making it central to cross-modal understanding. In Multimodal Continual Instruction Tuning (MCIT), however, shifting visual distributions and evolving instruction semantics cause this shared projector to drift, leading to projector-level forgetting, an issue largely overlooked by methods that focus primarily on the LLM backbone. We introduce Progressive Multimodal Alignment (PMA), a framework that enables the projector to adapt continually while preserving previously "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.26947","kind":"arxiv","version":1},"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/2607.26947/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-30T01:23:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ESVCk4cHK+5uSLBFK25WR9R6qYLQjfoN22ZEx6yijhct9sD7rCvGBxJJgjgZv1vKKXGaf1/jSzNavnr/L8/qCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T13:34:10.642604Z"},"content_sha256":"6ddd12c851442b5d3a39454d1298c19851b1fb207fed0516c8d0567098c9d53c","schema_version":"1.0","event_id":"sha256:6ddd12c851442b5d3a39454d1298c19851b1fb207fed0516c8d0567098c9d53c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ENR67QO2PL3NE43ESIDJWBQHOF/bundle.json","state_url":"https://pith.science/pith/ENR67QO2PL3NE43ESIDJWBQHOF/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ENR67QO2PL3NE43ESIDJWBQHOF/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-04T13:34:10Z","links":{"resolver":"https://pith.science/pith/ENR67QO2PL3NE43ESIDJWBQHOF","bundle":"https://pith.science/pith/ENR67QO2PL3NE43ESIDJWBQHOF/bundle.json","state":"https://pith.science/pith/ENR67QO2PL3NE43ESIDJWBQHOF/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ENR67QO2PL3NE43ESIDJWBQHOF/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:ENR67QO2PL3NE43ESIDJWBQHOF","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"7a5ed24acf989c9bfa0699d2bd4dd79d7ac028d7ee4c339aa4cabe64c170adb3","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-07-29T14:15:55Z","title_canon_sha256":"bf18220f5339ccbcd4236ed53c48f06c9bbed0db936970692051befeb40dc355"},"schema_version":"1.0","source":{"id":"2607.26947","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.26947","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"arxiv_version","alias_value":"2607.26947v1","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.26947","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"pith_short_12","alias_value":"ENR67QO2PL3N","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"pith_short_16","alias_value":"ENR67QO2PL3NE43E","created_at":"2026-07-30T01:23:08Z"},{"alias_kind":"pith_short_8","alias_value":"ENR67QO2","created_at":"2026-07-30T01:23:08Z"}],"graph_snapshots":[{"event_id":"sha256:6ddd12c851442b5d3a39454d1298c19851b1fb207fed0516c8d0567098c9d53c","target":"graph","created_at":"2026-07-30T01:23:08Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2607.26947/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Multimodal Large Language Models (MLLMs) rely on a projector to align visual representations with the language embedding space, making it central to cross-modal understanding. In Multimodal Continual Instruction Tuning (MCIT), however, shifting visual distributions and evolving instruction semantics cause this shared projector to drift, leading to projector-level forgetting, an issue largely overlooked by methods that focus primarily on the LLM backbone. We introduce Progressive Multimodal Alignment (PMA), a framework that enables the projector to adapt continually while preserving previously ","authors_text":"Duzhen Zhang, Jiahua Dong, Qiaoyi Su, Tielin Zhang, Yahan Yu","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-07-29T14:15:55Z","title":"Progressive Multimodal Alignment for Continual Instruction Tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.26947","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:4ffee3f3e02cd20034a3b766f55738e0b195876df3f09ae201352da7189292ca","target":"record","created_at":"2026-07-30T01:23:08Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"7a5ed24acf989c9bfa0699d2bd4dd79d7ac028d7ee4c339aa4cabe64c170adb3","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-07-29T14:15:55Z","title_canon_sha256":"bf18220f5339ccbcd4236ed53c48f06c9bbed0db936970692051befeb40dc355"},"schema_version":"1.0","source":{"id":"2607.26947","kind":"arxiv","version":1}},"canonical_sha256":"2363efc1da7af6d2736492069b060771679076928fcead58df7410722d8fc183","receipt":{"builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2363efc1da7af6d2736492069b060771679076928fcead58df7410722d8fc183","first_computed_at":"2026-07-30T01:23:08.181432Z","kind":"pith_receipt","last_reissued_at":"2026-07-30T01:23:08.181432Z","receipt_version":"0.3","signature_status":"unsigned_v0"},"source_id":"2607.26947","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4ffee3f3e02cd20034a3b766f55738e0b195876df3f09ae201352da7189292ca","sha256:6ddd12c851442b5d3a39454d1298c19851b1fb207fed0516c8d0567098c9d53c"],"state_sha256":"28c14dea1510d574841de4f812f231742fb31483e5373a1c7736b54e79945cdc"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Pe+8lKG974xZrCKR9R3rsNA1m4FqBFfiWs3r9WTIdv47zFdfOxBl3VXSNWATjigeiStiqvlVH2o5b+6GN+2VBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T13:34:10.647929Z","bundle_sha256":"fb7eb5a89d7072fb4c59050d6ad9e7bb795c1b636b9c655a57c70c0f7eeb40c1"}}