{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2601-19448","title":"From Internal Diagnosis to External Auditing: A VLM-Driven Paradigm for Data-Free Online Backdoor Defense","arxiv_id":"2601.19448","date":"2026-01-27","proceeding":"ICML","authors":["Binyan Xu","Fan Yang","Xilin Dai","Di Tang","Kehuan Zhang"],"abstract":"Deep Neural Networks remain inherently vulnerable to backdoor attacks. Traditional test-time defenses largely operate under the paradigm of internal diagnosis methods like model repairing or input robustness, yet these approaches are often fragile under advanced attacks as they remain entangled with the victim model's corrupted parameters. We propose a paradigm shift from Internal Diagnosis to External Semantic Auditing, arguing that effective defense requires decoupling safety from the victim model via an independent, semantically grounded auditor. To this end, we present a framework harnessing Universal Vision-Language Models (VLMs) as evolving semantic gatekeepers. We introduce PRISM (Prototype Refinement & Inspection via Statistical Monitoring), which overcomes the domain gap of general VLMs through two key mechanisms: a Hybrid VLM Teacher that dynamically refines visual prototypes online, and an Adaptive Router powered by statistical margin monitoring to calibrate gating thresholds in real-time. Extensive evaluation across 17 datasets and 11 attack types demonstrates that PRISM achieves state-of-the-art performance, suppressing Attack Success Rate to <1% on CIFAR-10 while improving clean accuracy, establishing a new standard for model-agnostic, externalized security.","url_abs":"https://arxiv.org/abs/2601.19448","url_pdf":"https://arxiv.org/pdf/2601.19448","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2601.19448","atlas_url":"https://app.syntology.ai/?focus=2601.19448","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2601.19448"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/binyxu/PRISM","reach":null}],"summary":{"unverified":3},"by_repo_kind":{"found_in_text":{"samples":3,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":3,"samples":[{"code_sha256_prefix":"0c07631b012f1cf9","entry":"fmt","repo":"binyxu/PRISM","repo_kind":"found_in_text","path":"rebuttal_experiments/reviewer_r3x4/W1_semantic_blend_attack/summarize_results.py","file_url":"https://github.com/binyxu/PRISM/blob/HEAD/rebuttal_experiments/reviewer_r3x4/W1_semantic_blend_attack/summarize_results.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0c07631b012f1cf9"}},{"code_sha256_prefix":"bc0ee5aebadb578f","entry":"get_attack_metrics","repo":"binyxu/PRISM","repo_kind":"found_in_text","path":"rebuttal_experiments/reviewer_r3x4/W1_semantic_blend_attack/summarize_results.py","file_url":"https://github.com/binyxu/PRISM/blob/HEAD/rebuttal_experiments/reviewer_r3x4/W1_semantic_blend_attack/summarize_results.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"bc0ee5aebadb578f"}},{"code_sha256_prefix":"af1a4b35c2c6a012","entry":"get_ca_asr_from_csv","repo":"binyxu/PRISM","repo_kind":"found_in_text","path":"rebuttal_experiments/reviewer_r3x4/W1_semantic_blend_attack/summarize_results.py","file_url":"https://github.com/binyxu/PRISM/blob/HEAD/rebuttal_experiments/reviewer_r3x4/W1_semantic_blend_attack/summarize_results.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"af1a4b35c2c6a012"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}