2026-09-19T07:53:22Z https://www.ideals.illinois.edu/oai-pmh

oai:www.ideals.illinois.edu:2142/132528 2026-02-20 col_2142_5131 col_2142_10761 com_2142_5130 com_2142_10755 com_2142_234

application/pdf en text Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2026-02-19 without embargo terms The student, Shubham Ugare, accepted the attached license on 2025-11-24 at 19:08. The student, Shubham Ugare, submitted this Dissertation for approval on 2025-11-24 at 19:14. This Dissertation was approved for publication on 2025-11-30 at 11:46. DSpace SAF Submission Ingestion Package generated from Vireo submission #22942 on 2026-02-19 at 18:25:17 Provably reliable machine learning systems Ugare, Shubham 2025-11-30 Singh, Gagandeep Misailovic, Sasa Zhang, Lingming Chaudhuri, Swarat Machine Learning, Formal Methods eng Machine learning systems, which primarily use deep neural networks (DNNs), serve as critical components in safety-critical applications and compound AI systems. Despite their ubiquity, automated formal reasoning about their reliability has lagged significantly. Neural network verification is NP-hard, requiring expensive end-to-end recomputation whenever networks are modified during iterative deployment cycles. Simultaneously, Large Language Models (LLMs) in compound systems frequently generate outputs that violate syntactic and semantic specifications, leading to cascading failures in automated workflows. Thus, developing reliability techniques for machine learning systems that are simultaneously general, precise, and scalable remains a challenging task. To address these challenges, this dissertation develops a comprehensive framework for provably reliable machine learning systems by establishing incremental verification techniques for DNNs and constrained decoding methods specific to LLMs. 2025-12 Thesis https://hdl.handle.net/2142/132528 Copyright 2025 Shubham Ugare Siebel School Comp & Data Sci Computer Science University of Illinois Urbana-Champaign Ph.D. Dissertation