{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FESBVAH3DFHLXE5F57DA34J3PX","short_pith_number":"pith:FESBVAH3","schema_version":"1.0","canonical_sha256":"29241a80fb194ebb93a5efc60df13b7dd1cb6a842b367cfda13963185ed09116","source":{"kind":"arxiv","id":"2505.21496","version":1},"attestation_state":"computed","paper":{"title":"UI-Genie: A Self-Improving Approach for Iteratively Boosting MLLM-based Mobile GUI Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aojun Zhou, Guozhi Wang, Han Xiao, Hao He, Hongsheng Li, Liang Liu, Liuyang Bian, Lue Fan, Rui Hu, Shuai Ren, Weifeng Lin, Xiaoxin Chen, Yafei Wen, Yuxiang Chai, Zimu Lu","submitted_at":"2025-05-27T17:58:06Z","abstract_excerpt":"In this paper, we introduce UI-Genie, a self-improving framework addressing two key challenges in GUI agents: verification of trajectory outcome is challenging and high-quality training data are not scalable. These challenges are addressed by a reward model and a self-improving pipeline, respectively. The reward model, UI-Genie-RM, features an image-text interleaved architecture that efficiently pro- cesses historical context and unifies action-level and task-level rewards. To sup- port the training of UI-Genie-RM, we develop deliberately-designed data genera- tion strategies including rule-ba"},"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":"2505.21496","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-27T17:58:06Z","cross_cats_sorted":["cs.CV","cs.LG"],"title_canon_sha256":"f4d6c55d21c89d096cf57527ecf84f5af41a0a77852479a8b4e04645bcc77fc1","abstract_canon_sha256":"9996739f14f9cb5f961ddebac64dd88b88e018f0909ce66f6cda143463342056"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:10:42.873767Z","signature_b64":"/8WNMaRg1P1GYeRMZPS9+8LGoksUa3VILDuni/IJ6UNQl+5d4XM07KP4SvS3qSzTwMwdJJYKIwIXaMfUT+z4AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"29241a80fb194ebb93a5efc60df13b7dd1cb6a842b367cfda13963185ed09116","last_reissued_at":"2026-07-05T11:10:42.873231Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:10:42.873231Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"UI-Genie: A Self-Improving Approach for Iteratively Boosting MLLM-based Mobile GUI Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aojun Zhou, Guozhi Wang, Han Xiao, Hao He, Hongsheng Li, Liang Liu, Liuyang Bian, Lue Fan, Rui Hu, Shuai Ren, Weifeng Lin, Xiaoxin Chen, Yafei Wen, Yuxiang Chai, Zimu Lu","submitted_at":"2025-05-27T17:58:06Z","abstract_excerpt":"In this paper, we introduce UI-Genie, a self-improving framework addressing two key challenges in GUI agents: verification of trajectory outcome is challenging and high-quality training data are not scalable. These challenges are addressed by a reward model and a self-improving pipeline, respectively. The reward model, UI-Genie-RM, features an image-text interleaved architecture that efficiently pro- cesses historical context and unifies action-level and task-level rewards. To sup- port the training of UI-Genie-RM, we develop deliberately-designed data genera- tion strategies including rule-ba"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.21496","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/2505.21496/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":"2505.21496","created_at":"2026-07-05T11:10:42.873300+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.21496v1","created_at":"2026-07-05T11:10:42.873300+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.21496","created_at":"2026-07-05T11:10:42.873300+00:00"},{"alias_kind":"pith_short_12","alias_value":"FESBVAH3DFHL","created_at":"2026-07-05T11:10:42.873300+00:00"},{"alias_kind":"pith_short_16","alias_value":"FESBVAH3DFHLXE5F","created_at":"2026-07-05T11:10:42.873300+00:00"},{"alias_kind":"pith_short_8","alias_value":"FESBVAH3","created_at":"2026-07-05T11:10:42.873300+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.11078","citing_title":"A History-Aware Visually Grounded Critic for Computer Use Agents","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14311","citing_title":"Beyond Binary: Reframing GUI Critique as Continuous Semantic Alignment","ref_index":118,"is_internal_anchor":false},{"citing_arxiv_id":"2512.12634","citing_title":"MobiBench: Multi-Branch, Modular Benchmark for Mobile GUI Agents","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14311","citing_title":"Beyond Binary: Reframing GUI Critique as Continuous Semantic Alignment","ref_index":118,"is_internal_anchor":false},{"citing_arxiv_id":"2507.21046","citing_title":"A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence","ref_index":156,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22558","citing_title":"SOLAR-RL: Semi-Online Long-horizon Assignment Reinforcement Learning","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10674","citing_title":"Skill-SD: Skill-Conditioned Self-Distillation for Multi-turn LLM Agents","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07110","citing_title":"Securing Computer-Use Agents: A Unified Architecture-Lifecycle Framework for Deployment-Grounded Reliability","ref_index":116,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX","json":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX.json","graph_json":"https://pith.science/api/pith-number/FESBVAH3DFHLXE5F57DA34J3PX/graph.json","events_json":"https://pith.science/api/pith-number/FESBVAH3DFHLXE5F57DA34J3PX/events.json","paper":"https://pith.science/paper/FESBVAH3"},"agent_actions":{"view_html":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX","download_json":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX.json","view_paper":"https://pith.science/paper/FESBVAH3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.21496&json=true","fetch_graph":"https://pith.science/api/pith-number/FESBVAH3DFHLXE5F57DA34J3PX/graph.json","fetch_events":"https://pith.science/api/pith-number/FESBVAH3DFHLXE5F57DA34J3PX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX/action/storage_attestation","attest_author":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX/action/author_attestation","sign_citation":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX/action/citation_signature","submit_replication":"https://pith.science/pith/FESBVAH3DFHLXE5F57DA34J3PX/action/replication_record"}},"created_at":"2026-07-05T11:10:42.873300+00:00","updated_at":"2026-07-05T11:10:42.873300+00:00"}