CORTEXA
← Browse
arxivcs.AIcs.LG2026-06-30

Self-Study Reconsidered: The Hidden Fragility of Learning from Self-Generated QA

Ekaterina Alimaskina, Denis Shveykin, Gleb Molodtsov, Igor Shalygin, Alexey Kadeishvili, Aleksandr Beznosikov

Language models are increasingly taught from synthetic question--answer (QA) supervision: a model generates questions about a document, answers them from the same text, and the resulting pairs are used to fine-tune, distill, or compress knowledge into another model. We show that this generation step is not neutral preprocessing. It is an implicit policy that both selects which evidence becomes training signal and decides how that evidence is answered, and it is fragile at both stages. When choosing what to ask, generators do not scan a document uniformly. Coverage saturates early and concentrates on salient spans, diverse prompts converge on the same regions, and what looks question-worthy is driven by local presentation. As a result, salient artifacts such as poorly cleaned markup can hijack question generation across model families and scales. When answering, the model that produces the supervision tends to obey instruction-like passages embedded in the text. This compliance depends on the intent and surface form of the passage rather than its strictness, and is worst under task conflict, where larger models comply more often. These failure modes arise from choices made during QA generation, so they can be reduced without changing the training loop. Tying each question to a fixed target reduces biased selection, and filtering instruction-like spans before answering lowers mean injection compliance from $88\%$ to $13\%$ in our evaluation while retaining nearly all clean text.

View free PDFSource page

Related papers

arxivcs.AIcs.LG2026-07-15

CatalogAgent: A Supervisor-mediated Self-Learning System Enabling Context Engineering for GenAI Models

Zhu Cheng, Zhenming Wang, Yu, Tang, Dan Liu, Bryan Zhang, et al.

Product catalogs are the backbone of e-commerce sites, yet a large number of structured attributes (SAs) -- such as material, color, and shape -- often have missing values. Typically, SA values are extracted from product information, including titles and descriptions. While LLM-b…

View free PDFSource page
arxivcs.LGcs.AI2026-07-21Cited by 18

Breaking the Homogeneity Assumption: Specialized Multi-Generator Adversarial Learning for Rare Failure Detection in Predictive Maintenance

Alexis Lazanas, Georgios Kampouropoulos

Supervised learning models in the predictive maintenance field are regularly trained on highly imbalanced industrial datasets: machine failures occur rarely but have a disproportionate effect on operations. In addition to the clear class disparity, failure data are typically non-…

View free PDFSource page
arxivcs.LGcs.AI2026-07-20

Physical Self-Supervised Learning: IMU Sensing without Manual Labels

Yuyang Leng, Renyuan Liu, Shaohan Hu, Peijun Zhao, Chun-Fu Chen, Songqing Chen, et al.

Deep neural networks have become a promising approach for IMU-based sensing, but their scalability is fundamentally limited by costly labeled data and poor robustness to heterogeneous devices, placements, and users. Existing unsupervised and self-supervised methods reduce but do…

View free PDFSource page
arxivcs.LGcs.AI2026-07-06

Self-Review Reinforcement Learning (SRRL) with Cross-Episode Memory and Policy Distillation

Muhammad Zain Amin, Kibele Sebnem Yildirim

Reinforcement Learning is commonly used to train large language models using environmental feedback. In applied settings, the environment usually provides sparse or delayed feedback. This makes it difficult for the model to pinpoint which actions in its reasoning led to success o…

View free PDFSource page