{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6OGYIGJ35F5ZOCQ42PJRUH7PKH","short_pith_number":"pith:6OGYIGJ3","schema_version":"1.0","canonical_sha256":"f38d84193be97b970a1cd3d31a1fef51eabb49fdd6b6aa1ab3dfa52fb601c815","source":{"kind":"arxiv","id":"2606.17871","version":1},"attestation_state":"computed","paper":{"title":"StepGuard: Guarding Web Navigation via Single-Step Calibration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jinpeng Hu, Liu Liu, Li Zhu, Mengjia Li, Xiyang Sun, Yaxiong Wang, Yuchen Zhang, Yujiao Wu, Zhihao Cui","submitted_at":"2026-06-16T12:42:09Z","abstract_excerpt":"Web navigation requires agents to follow natural language goals, interact with web pages, and produce accurate answers. While recent advances leverage vision-language models and reinforcement learning, existing methods still suffer from single-step fragility due to reward misalignment and error propagation. To tackle the reward entanglement, we design Dynamic Dual-Policy Optimization (DDPO), which dynamically switches between a navigation-first mode for exploration and an answer-first mode for question-answering to mitigate reward conflict. To calibrate the single-step error, we propose Confid"},"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":"2606.17871","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-16T12:42:09Z","cross_cats_sorted":[],"title_canon_sha256":"06f537d9742a7c587460202be0123d23854b17d5fbdeab6189165e3c9485ccc3","abstract_canon_sha256":"667959f78736fee0cff85b2f26119f89473f4ae8d7a1c08f0986554a899dedb4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:10:42.365785Z","signature_b64":"+V732VW1RftcU0RPip+emUDw1pKdchOpx/IvIYF952yF17MEuR2xpdxRClAonp/5RE7WCO6uwl4pNUXJ9X8qAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f38d84193be97b970a1cd3d31a1fef51eabb49fdd6b6aa1ab3dfa52fb601c815","last_reissued_at":"2026-06-19T16:10:42.365376Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:10:42.365376Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"StepGuard: Guarding Web Navigation via Single-Step Calibration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jinpeng Hu, Liu Liu, Li Zhu, Mengjia Li, Xiyang Sun, Yaxiong Wang, Yuchen Zhang, Yujiao Wu, Zhihao Cui","submitted_at":"2026-06-16T12:42:09Z","abstract_excerpt":"Web navigation requires agents to follow natural language goals, interact with web pages, and produce accurate answers. While recent advances leverage vision-language models and reinforcement learning, existing methods still suffer from single-step fragility due to reward misalignment and error propagation. To tackle the reward entanglement, we design Dynamic Dual-Policy Optimization (DDPO), which dynamically switches between a navigation-first mode for exploration and an answer-first mode for question-answering to mitigate reward conflict. To calibrate the single-step error, we propose Confid"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.17871","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/2606.17871/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":"2606.17871","created_at":"2026-06-19T16:10:42.365434+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.17871v1","created_at":"2026-06-19T16:10:42.365434+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.17871","created_at":"2026-06-19T16:10:42.365434+00:00"},{"alias_kind":"pith_short_12","alias_value":"6OGYIGJ35F5Z","created_at":"2026-06-19T16:10:42.365434+00:00"},{"alias_kind":"pith_short_16","alias_value":"6OGYIGJ35F5ZOCQ4","created_at":"2026-06-19T16:10:42.365434+00:00"},{"alias_kind":"pith_short_8","alias_value":"6OGYIGJ3","created_at":"2026-06-19T16:10:42.365434+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH","json":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH.json","graph_json":"https://pith.science/api/pith-number/6OGYIGJ35F5ZOCQ42PJRUH7PKH/graph.json","events_json":"https://pith.science/api/pith-number/6OGYIGJ35F5ZOCQ42PJRUH7PKH/events.json","paper":"https://pith.science/paper/6OGYIGJ3"},"agent_actions":{"view_html":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH","download_json":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH.json","view_paper":"https://pith.science/paper/6OGYIGJ3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.17871&json=true","fetch_graph":"https://pith.science/api/pith-number/6OGYIGJ35F5ZOCQ42PJRUH7PKH/graph.json","fetch_events":"https://pith.science/api/pith-number/6OGYIGJ35F5ZOCQ42PJRUH7PKH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH/action/storage_attestation","attest_author":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH/action/author_attestation","sign_citation":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH/action/citation_signature","submit_replication":"https://pith.science/pith/6OGYIGJ35F5ZOCQ42PJRUH7PKH/action/replication_record"}},"created_at":"2026-06-19T16:10:42.365434+00:00","updated_at":"2026-06-19T16:10:42.365434+00:00"}