Automatically Detecting Numerical Instability in Machine Learning Applications via Soft Assertions

• URL: http://arxiv.org/abs/2504.15507v2
• Date: Wed, 23 Apr 2025 07:46:21 GMT
• Title: Automatically Detecting Numerical Instability in Machine Learning Applications via Soft Assertions
• Authors: Shaila Sharmin, Anwar Hossain Zahid, Subhankar Bhattacharjee, Chiamaka Igwilo, Miryung Kim, Wei Le,
• Abstract summary: numerical bugs can lead to system crashes, incorrect output, and wasted computing resources.<n>We introduce a novel idea, namely soft assertions (SA), to encode safety/error conditions for the places where numerical instability can occur.
• Score: 7.893728124841138
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Machine learning (ML) applications have become an integral part of our lives. ML applications extensively use floating-point computation and involve very large/small numbers; thus, maintaining the numerical stability of such complex computations remains an important challenge. Numerical bugs can lead to system crashes, incorrect output, and wasted computing resources. In this paper, we introduce a novel idea, namely soft assertions (SA), to encode safety/error conditions for the places where numerical instability can occur. A soft assertion is an ML model automatically trained using the dataset obtained during unit testing of unstable functions. Given the values at the unstable function in an ML application, a soft assertion reports how to change these values in order to trigger the instability. We then use the output of soft assertions as signals to effectively mutate inputs to trigger numerical instability in ML applications. In the evaluation, we used the GRIST benchmark, a total of 79 programs, as well as 15 real-world ML applications from GitHub. We compared our tool with 5 state-of-the-art (SOTA) fuzzers. We found all the GRIST bugs and outperformed the baselines. We found 13 numerical bugs in real-world code, one of which had already been confirmed by the GitHub developers. While the baselines mostly found the bugs that report NaN and INF, our tool \tool found numerical bugs with incorrect output. We showed one case where the Tumor Detection Model, trained on Brain MRI images, should have predicted "tumor", but instead, it incorrectly predicted "no tumor" due to the numerical bugs. Our replication package is located at https://figshare.com/s/6528d21ccd28bea94c32.

Related papers

• MICE for CATs: Model-Internal Confidence Estimation for Calibrating Agents with Tools [54.63478102768333]
  Well-calibrated model confidences can be used to weigh the risk versus reward of potential actions. We propose a novel class of model-internal confidence estimators (MICE) to better assess confidence when calling tools.
  arXiv  Detail & Related papers  (2025-04-28T18:06:38Z)
• Subgraph-Oriented Testing for Deep Learning Libraries [9.78188667672054]
  We propose SORT (Subgraph-Oriented Realistic Testing) to test Deep Learning (DL) libraries on different hardware platforms.<n>SORT takes popular API interaction patterns, represented as frequent subgraphs of model graphs, as test subjects.<n>SORT achieves a 100% valid input generation rate, detects more precision bugs than existing methods, and reveals interaction-related bugs missed by single-API testing.
  arXiv  Detail & Related papers  (2024-12-09T12:10:48Z)
• Fast Fixes and Faulty Drivers: An Empirical Analysis of Regression Bug Fixing Times in the Linux Kernel [3.1959458747110054]
  The paper focuses on regression bug tracking in the kernel by considering the time required to fix regression bugs. The dataset examined is based on the regzbot automation framework for tracking regressions in the Linux kernel.
  arXiv  Detail & Related papers  (2024-11-04T13:53:29Z)
• KGym: A Platform and Dataset to Benchmark Large Language Models on Linux Kernel Crash Resolution [59.20933707301566]
  Large Language Models (LLMs) are consistently improving at increasingly realistic software engineering (SE) tasks. In real-world software stacks, significant SE effort is spent developing foundational system software like the Linux kernel. To evaluate if ML models are useful while developing such large-scale systems-level software, we introduce kGym and kBench.
  arXiv  Detail & Related papers  (2024-07-02T21:44:22Z)
• Neuron Patching: Semantic-based Neuron-level Language Model Repair for Code Generation [32.178931149612644]
  ulModel ulImprovement via ulNeuron ulTargeting (textscMINT) is a novel approach for repairing code Language Models (LMs) textscMINT is effective, efficient, and reliable, capable of correcting a neural model by patching a minimum number of neurons.
  arXiv  Detail & Related papers  (2023-12-08T20:28:08Z)
• ML-Bench: Evaluating Large Language Models and Agents for Machine Learning Tasks on Repository-Level Code [76.84199699772903]
  ML-Bench is a benchmark rooted in real-world programming applications that leverage existing code repositories to perform tasks. To evaluate both Large Language Models (LLMs) and AI agents, two setups are employed: ML-LLM-Bench for assessing LLMs' text-to-code conversion within a predefined deployment environment, and ML-Agent-Bench for testing autonomous agents in an end-to-end task execution within a Linux sandbox environment.
  arXiv  Detail & Related papers  (2023-11-16T12:03:21Z)
• Test-Time Self-Adaptive Small Language Models for Question Answering [63.91013329169796]
  We show and investigate the capabilities of smaller self-adaptive LMs, only with unlabeled test data. Our proposed self-adaption strategy demonstrates significant performance improvements on benchmark QA datasets.
  arXiv  Detail & Related papers  (2023-10-20T06:49:32Z)
• Guess & Sketch: Language Model Guided Transpilation [59.02147255276078]
  Learned transpilation offers an alternative to manual re-writing and engineering efforts. Probabilistic neural language models (LMs) produce plausible outputs for every input, but do so at the cost of guaranteed correctness. Guess & Sketch extracts alignment and confidence information from features of the LM then passes it to a symbolic solver to resolve semantic equivalence.
  arXiv  Detail & Related papers  (2023-09-25T15:42:18Z)
• Fuzzing with Quantitative and Adaptive Hot-Bytes Identification [6.442499249981947]
  American fuzzy lop, a leading fuzzing tool, has demonstrated its powerful bug finding ability through a vast number of reported CVEs. We propose an approach called toolwhich is designed based on the following principles. Our evaluation results on 10 real-world programs and LAVA-M dataset show that toolachieves sustained increases in branch coverage and discovers more bugs than other fuzzers.
  arXiv  Detail & Related papers  (2023-07-05T13:41:35Z)
• Fault-Aware Neural Code Rankers [64.41888054066861]
  We propose fault-aware neural code rankers that can predict the correctness of a sampled program without executing it. Our fault-aware rankers can significantly increase the pass@1 accuracy of various code generation models.
  arXiv  Detail & Related papers  (2022-06-04T22:01:05Z)
• Vamsa: Automated Provenance Tracking in Data Science Scripts [17.53546311589593]
  We introduce the ML provenance tracking problem. We discuss the challenges in capturing such information in the context of Python. We present Vamsa, a modular system that extracts provenance from Python scripts without requiring any changes to the users' code.
  arXiv  Detail & Related papers  (2020-01-07T02:39:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.