{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:HRECSQNP76GOLNUB7X2HOBXYDA","short_pith_number":"pith:HRECSQNP","schema_version":"1.0","canonical_sha256":"3c482941afff8ce5b681fdf47706f8181588bf9ce52f4d3987154b1311c4f0c5","source":{"kind":"arxiv","id":"1901.10902","version":5},"attestation_state":"computed","paper":{"title":"InfoBot: Transfer and Exploration via the Information Bottleneck","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Anirudh Goyal, Daniel Strouse, Hugo Larochelle, Matthew Botvinick, Riashat Islam, Sergey Levine, Yoshua Bengio, Zafarali Ahmed","submitted_at":"2019-01-30T15:33:58Z","abstract_excerpt":"A central challenge in reinforcement learning is discovering effective policies for tasks where rewards are sparsely distributed. We postulate that in the absence of useful reward signals, an effective exploration strategy should seek out {\\it decision states}. These states lie at critical junctions in the state space from where the agent can transition to new, potentially unexplored regions. We propose to learn about decision states from prior experience. By training a goal-conditioned policy with an information bottleneck, we can identify decision states by examining where the model actually"},"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":"1901.10902","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2019-01-30T15:33:58Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"0657b8a0677e0f10b9df92e4ac51bbfaf83cb2b1f6bf2b0845a87770131bb197","abstract_canon_sha256":"bea6f0ffb80495281b92a118b739d50796bbef94cbcbd09d7d48af71858d339a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:20:46.013016Z","signature_b64":"CbhgPWQ3dJqkCBQDTx/7u5VnTbvtGPP1JxMXjdhACeWhbsyClM/f1KOQ94OMq99Z6aaCKNkO3+ce1WAeOObFAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3c482941afff8ce5b681fdf47706f8181588bf9ce52f4d3987154b1311c4f0c5","last_reissued_at":"2026-07-05T07:20:46.012515Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:20:46.012515Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"InfoBot: Transfer and Exploration via the Information Bottleneck","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Anirudh Goyal, Daniel Strouse, Hugo Larochelle, Matthew Botvinick, Riashat Islam, Sergey Levine, Yoshua Bengio, Zafarali Ahmed","submitted_at":"2019-01-30T15:33:58Z","abstract_excerpt":"A central challenge in reinforcement learning is discovering effective policies for tasks where rewards are sparsely distributed. We postulate that in the absence of useful reward signals, an effective exploration strategy should seek out {\\it decision states}. These states lie at critical junctions in the state space from where the agent can transition to new, potentially unexplored regions. We propose to learn about decision states from prior experience. By training a goal-conditioned policy with an information bottleneck, we can identify decision states by examining where the model actually"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1901.10902","kind":"arxiv","version":5},"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/1901.10902/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":"1901.10902","created_at":"2026-07-05T07:20:46.012576+00:00"},{"alias_kind":"arxiv_version","alias_value":"1901.10902v5","created_at":"2026-07-05T07:20:46.012576+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1901.10902","created_at":"2026-07-05T07:20:46.012576+00:00"},{"alias_kind":"pith_short_12","alias_value":"HRECSQNP76GO","created_at":"2026-07-05T07:20:46.012576+00:00"},{"alias_kind":"pith_short_16","alias_value":"HRECSQNP76GOLNUB","created_at":"2026-07-05T07:20:46.012576+00:00"},{"alias_kind":"pith_short_8","alias_value":"HRECSQNP","created_at":"2026-07-05T07:20:46.012576+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19998","citing_title":"Tri-Info: Generalizable, Interpretable Failure Prediction for VLA Models via Information Theory","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27532","citing_title":"SCALE-COMM: Shared, Contrastively-Aligned Latent Embeddings for MARL Communication","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12236","citing_title":"TMRL: Diffusion Timestep-Modulated Pretraining Enables Exploration for Efficient Policy Finetuning","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08526","citing_title":"Skill-CMIB: Multimodal Agent Skill for Consistent Action via Conditional Multimodal Information Bottleneck","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18460","citing_title":"Learning Invariant Modality Representation for Robust Multimodal Learning from a Causal Inference Perspective","ref_index":120,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA","json":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA.json","graph_json":"https://pith.science/api/pith-number/HRECSQNP76GOLNUB7X2HOBXYDA/graph.json","events_json":"https://pith.science/api/pith-number/HRECSQNP76GOLNUB7X2HOBXYDA/events.json","paper":"https://pith.science/paper/HRECSQNP"},"agent_actions":{"view_html":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA","download_json":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA.json","view_paper":"https://pith.science/paper/HRECSQNP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1901.10902&json=true","fetch_graph":"https://pith.science/api/pith-number/HRECSQNP76GOLNUB7X2HOBXYDA/graph.json","fetch_events":"https://pith.science/api/pith-number/HRECSQNP76GOLNUB7X2HOBXYDA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA/action/storage_attestation","attest_author":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA/action/author_attestation","sign_citation":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA/action/citation_signature","submit_replication":"https://pith.science/pith/HRECSQNP76GOLNUB7X2HOBXYDA/action/replication_record"}},"created_at":"2026-07-05T07:20:46.012576+00:00","updated_at":"2026-07-05T07:20:46.012576+00:00"}