Debugging Machine Learning Tasks

By Aleksandar Chakarov, Aditya V. Nori, Sriram K. Rajamani, Shayak Sen, Deepak Vijaykeerthy. 2016

47 citations recorded by the source corpus.

Unlike traditional programs (such as operating systems or word processors) which have large amounts of code, machine learning tasks use programs with relatively small amounts of code (written in machine learning libraries), but voluminous amounts of data. Just like developers of traditional programs debug errors in their code, developers of machine learning tasks debug and fix errors in their data. However, algorithms and tools for debugging and fixing errors in data are less common, when compared to their counterparts for detecting and fixing errors in code. In this paper, we consider classification tasks where errors in training data lead to misclassifications in test points, and propose an automated method to find the root causes of such misclassifications. Our root cause analysis is based on Pearl's theory of causation, and uses Pearl's PS (Probability of Sufficiency) as a scoring metric. Our implementation, Psi, encodes the computation of PS as a probabilistic program, and uses recent work on probabilistic programs and transformations on probabilistic programs (along with gray-box models of machine learning algorithms) to efficiently compute PS. Psi is able to identify root causes of data errors in interesting data sets.

Subfields: Artificial Intelligence, Information Systems, Software

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Software

Citations in the corpus

Generative Models for Effective ML on Private, Decentralized Datasets (2019, background): “In general, an assumption of unfettered access to training or inference data is made (Riley, 2016; Zinkevich, 2017; Chakarov et al., 2016).”
On the Safety of Machine Learning: Cyber-Physical Systems, Decision Sciences, and Data Products (2016, neutral): “No competing financial interests exist.”
Mitigating Poisoning Attacks on Machine Learning Models: A Data Provenance Based Approach (2017, methodology): “Two prior methods, Reject on Negative Impact (RONI) [13] and the Probability of Sufficiency (PS) method by [7], take a similar approach of detecting poison data by evaluating the effect of individual data points on the performance of the trained model.”
A survey of deep learning techniques for autonomous driving (2019, background): “s have been cataloged for neural networks in (Kurd et al., 2007), (Harris, 2016) and (McPherson, 2018). This led to the development of speciﬁc and focused tools and techniques to help ﬁnding faults. (Chakarov et al., 2018) describes a technique for debugging misclassiﬁcations due to bad training dat”
Using Machine Learning Safely in Automotive Software: An Assessment and Adaption of Software Process Requirements in ISO 26262 (2018, methodology): “[8] describe a technique for debugging misclassifications due to bad training data, while Nushi et al.”
Debugging Tests for Model Explanations (2020, background): “Diagnosing and ﬁxing model errors–model debugging–remains a longstanding machine learning challenge [12, 14–17, 55, 73].”
On Human Intellect and Machine Failures: Troubleshooting Integrative Machine Learning Systems (2016, methodology): “On debugging a single classifier, researchers have developed techniques for a domain expert to interact with the machine learning process [4, 9, 13, 1].”