{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4NPPWFDSH7M666K2EPYUXTCUYW","short_pith_number":"pith:4NPPWFDS","schema_version":"1.0","canonical_sha256":"e35efb14723fd9ef795a23f14bcc54c58cd50bf6e7652cb810f728f00e33b9d5","source":{"kind":"arxiv","id":"2506.17218","version":1},"attestation_state":"computed","paper":{"title":"Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chuang Gan, Delin Chen, Maohao Shen, Xueyang Yu, Zeyuan Yang","submitted_at":"2025-06-20T17:59:31Z","abstract_excerpt":"Vision-language models (VLMs) excel at multimodal understanding, yet their text-only decoding forces them to verbalize visual reasoning, limiting performance on tasks that demand visual imagination. Recent attempts train VLMs to render explicit images, but the heavy image-generation pre-training often hinders the reasoning ability. Inspired by the way humans reason with mental imagery-the internal construction and manipulation of visual cues-we investigate whether VLMs can reason through interleaved multimodal trajectories without producing explicit images. To this end, we present a Machine Me"},"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":"2506.17218","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-06-20T17:59:31Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e0de502ca162d4f3dafd62f4e4c966d0cd99ddd23f6b77b505e00432c15ea88e","abstract_canon_sha256":"dae1cfd5ed21e534c7c869b9ad17b1dbd21cd28467b078177ac3e317c33e4410"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:47.204908Z","signature_b64":"x9ZzK0HW9j1VtDb+cqTWAOyyKgw/uOWV9HaFJ3PVndjE67iFlS6HnU75eJhasuKd7oz9lIyMkwefpF0WwHsNBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e35efb14723fd9ef795a23f14bcc54c58cd50bf6e7652cb810f728f00e33b9d5","last_reissued_at":"2026-05-17T23:38:47.203851Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:47.203851Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chuang Gan, Delin Chen, Maohao Shen, Xueyang Yu, Zeyuan Yang","submitted_at":"2025-06-20T17:59:31Z","abstract_excerpt":"Vision-language models (VLMs) excel at multimodal understanding, yet their text-only decoding forces them to verbalize visual reasoning, limiting performance on tasks that demand visual imagination. Recent attempts train VLMs to render explicit images, but the heavy image-generation pre-training often hinders the reasoning ability. Inspired by the way humans reason with mental imagery-the internal construction and manipulation of visual cues-we investigate whether VLMs can reason through interleaved multimodal trajectories without producing explicit images. To this end, we present a Machine Me"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.17218","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":""},"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":"2506.17218","created_at":"2026-05-17T23:38:47.203945+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.17218v1","created_at":"2026-05-17T23:38:47.203945+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.17218","created_at":"2026-05-17T23:38:47.203945+00:00"},{"alias_kind":"pith_short_12","alias_value":"4NPPWFDSH7M6","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"4NPPWFDSH7M666K2","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"4NPPWFDS","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":27,"internal_anchor_count":5,"sample":[{"citing_arxiv_id":"2509.23108","citing_title":"Artificial Phantasia: Emergent Mental Imagery in Large Language Models","ref_index":85,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12374","citing_title":"Fill the GAP: A Granular Alignment Paradigm for Visual Reasoning in Multimodal Large Language Models","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19342","citing_title":"Semantic-Enriched Latent Visual Reasoning","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12374","citing_title":"Fill the GAP: A Granular Alignment Paradigm for Visual Reasoning in Multimodal Large Language Models","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2512.12623","citing_title":"Reasoning Within the Mind: Dynamic Multimodal Interleaving in Latent Space","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2601.06803","citing_title":"Forest Before Trees: Latent Superposition for Efficient Visual Reasoning","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2602.11731","citing_title":"Thinking with Drafting: Optical Decompression via Logical Reconstruction","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2602.09850","citing_title":"Towards Explainable Industrial Anomaly Detection via Knowledge-Guided Latent Reasoning","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12163","citing_title":"Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Model","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03307","citing_title":"V-Reflection: Transforming MLLMs from Passive Observers to Active Interrogators","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03179","citing_title":"Understanding the Role of Hallucination in Reinforcement Post-Training of Multimodal Reasoning Models","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11856","citing_title":"UniVLR: Unifying Text and Vision in Visual Latent Reasoning for Multimodal LLMs","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10500","citing_title":"Visual Enhanced Depth Scaling for Multimodal Latent Reasoning","ref_index":71,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12374","citing_title":"Fill the GAP: A Granular Alignment Paradigm for Visual Reasoning in Multimodal Large Language Models","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12163","citing_title":"Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Model","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08802","citing_title":"CoLVR: Enhancing Exploratory Latent Visual Reasoning via Contrastive Optimization","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.28192","citing_title":"LaST-R1: Reinforcing Robotic Manipulation via Adaptive Physical Latent Reasoning","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08802","citing_title":"CoLVR: Enhancing Exploratory Latent Visual Reasoning via Contrastive Optimization","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10500","citing_title":"Visual Enhanced Depth Scaling for Multimodal Latent Reasoning","ref_index":71,"is_internal_anchor":false},{"citing_arxiv_id":"2604.28192","citing_title":"LaST-R1: Reinforcing Robotic Manipulation via Adaptive Physical Latent Reasoning","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02735","citing_title":"Visual Latents Know More Than They Say: Unsilencing Latent Reasoning in MLLMs","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20328","citing_title":"Hybrid Latent Reasoning with Decoupled Policy Optimization","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12896","citing_title":"Don't Show Pixels, Show Cues: Unlocking Visual Tool Reasoning in Language Models via Perception Programs","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10500","citing_title":"Visual Enhanced Depth Scaling for Multimodal Latent Reasoning","ref_index":71,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07518","citing_title":"Decompose, Look, and Reason: Reinforced Latent Reasoning for VLMs","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW","json":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW.json","graph_json":"https://pith.science/api/pith-number/4NPPWFDSH7M666K2EPYUXTCUYW/graph.json","events_json":"https://pith.science/api/pith-number/4NPPWFDSH7M666K2EPYUXTCUYW/events.json","paper":"https://pith.science/paper/4NPPWFDS"},"agent_actions":{"view_html":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW","download_json":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW.json","view_paper":"https://pith.science/paper/4NPPWFDS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.17218&json=true","fetch_graph":"https://pith.science/api/pith-number/4NPPWFDSH7M666K2EPYUXTCUYW/graph.json","fetch_events":"https://pith.science/api/pith-number/4NPPWFDSH7M666K2EPYUXTCUYW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW/action/storage_attestation","attest_author":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW/action/author_attestation","sign_citation":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW/action/citation_signature","submit_replication":"https://pith.science/pith/4NPPWFDSH7M666K2EPYUXTCUYW/action/replication_record"}},"created_at":"2026-05-17T23:38:47.203945+00:00","updated_at":"2026-05-17T23:38:47.203945+00:00"}