Training Data Influence Analysis and Estimation: A Survey

Hammoudeh, Zayd; Lowd, Daniel

doi:10.48550/arxiv.2212.04612

Cited by 3 publications

(7 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Training data play a fundamental role in the success of AI algorithms, as they provide a set of cases or instances to the model to teach it to perform a specific task [ 21 , 22 ]. These data consist of a combination of expected inputs and outputs, so that the model learns to associate the inputs with their respective correct outputs from a training process through functions that capture the dynamics of the system; this process is called training [ 22 , 23 ].…”

Section: Training Data—data Setmentioning

confidence: 99%

“…It is essential to ensure that the data are sufficient, representative, relevant, unrelated, and diverse, as this will allow the model to capture the complexity (generalization) of the scenarios in which AI will be used. Additionally, it is crucial to have the data correctly labeled and annotated so that the models can learn effectively [ 21 ].…”

Section: Training Data—data Setmentioning

confidence: 99%

See 1 more Smart Citation

Machine Learning to Predict Enzyme–Substrate Interactions in Elucidation of Synthesis Pathways: A Review

Salas-Nuñez,

Barrera-Ocampo,

Caicedo

et al. 2024

Metabolites

View full text Add to dashboard Cite

Enzyme–substrate interactions play a fundamental role in elucidating synthesis pathways and synthetic biology, as they allow for the understanding of important aspects of a reaction. Establishing the interaction experimentally is a slow and costly process, which is why this problem has been addressed using computational methods such as molecular dynamics, molecular docking, and Monte Carlo simulations. Nevertheless, this type of method tends to be computationally slow when dealing with a large search space. Therefore, in recent years, methods based on artificial intelligence, such as support vector machines, neural networks, or decision trees, have been implemented, significantly reducing the computing time and covering vast search spaces. These methods significantly reduce the computation time and cover broad search spaces, rapidly reducing the number of interacting candidates, as they allow repetitive processes to be automated and patterns to be extracted, are adaptable, and have the capacity to handle large amounts of data. This article analyzes these artificial intelligence-based approaches, presenting their common structure, advantages, disadvantages, limitations, challenges, and future perspectives.

show abstract

Section: Training Data—data Setmentioning

confidence: 99%

Section: Training Data—data Setmentioning

confidence: 99%

Machine Learning to Predict Enzyme–Substrate Interactions in Elucidation of Synthesis Pathways: A Review

Salas-Nuñez,

Barrera-Ocampo,

Caicedo

et al. 2024

Metabolites

View full text Add to dashboard Cite

show abstract

“…Doshi-Velez and Kim [8] stress the necessity for evaluation techniques for XAI, while Carvalho et al [9] offer an extensive study of interpretability methods covering model-agnostic and model-specific approaches. Hammoudeh and Lowd [174] shift the focus to the influence of training data. Mohseni et al provide a survey and framework to evaluate XAI systems [175].…”

Section: ) Data Preparation and Transformationmentioning

confidence: 99%

“…Interpreting deep models by measuring the influence of training samples on decision making is essential for understanding and validating the outputs of these models. The process generally involves several techniques that map the correlation between individual training samples and the decisions made by the model [221], [174]. In this section, we categorize existing work into three folders as follows.…”

Section: Influences: Data Valuation and Anomaly Detection Of Training...mentioning

confidence: 99%

“…The comprehensive nature of these works bridges the gap between theory and real-world applications. Meanwhile, Hammoudeh and Lowd [174] provide a distinct perspective, shifting the attention from model interpretability to the role of training data in models. They detail the impacts and various methods for determining the importance of training data contributions.…”

Section: A1 Related Surveysmentioning

confidence: 99%

See 1 more Smart Citation

Classification of Kinematic Data Using Explainable Artificial Intelligence (XAI) for Smart Motion

Yong¹

2021

Explainable Artificial Intelligence for Smart Cities

View full text Add to dashboard Cite

Given the complexity and lack of transparency in deep neural networks (DNNs), extensive efforts have been made to make these systems more interpretable or explain their behaviors in accessible terms. Unlike most reviews, which focus on algorithmic and model-centric perspectives, this work takes a "data-centric" view, examining how data collection, processing, and analysis contribute to explainable AI (XAI). We categorize existing work into three categories subject to their purposes: interpretations of deep models, referring to feature attributions and reasoning processes that correlate data points with model outputs; influences of training data, examining the impact of training data nuances, such as data valuation and sample anomalies, on decision-making processes; and insights of domain knowledge, discovering latent patterns and fostering new knowledge from data and models to advance social values and scientific discovery. Specifically, we distill XAI methodologies into data mining operations on training and testing data across modalities, such as images, text, and tabular data, as well as on training logs, checkpoints, models and other DNN behavior descriptors. In this way, our study offers a comprehensive, data-centric examination of XAI from a lens of data mining methods and applications.

show abstract

Holistic evaluation of a machine learning-based timing calibration for PET detectors under varying data sparsity

Naunheim,

Mueller,

Nadig

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

Objective.Modern PET scanners offer precise TOF information, improving the SNR of the reconstructed images. Timing calibrations are performed to reduce the worsening effects of the system components and provide valuable TOF information. Traditional calibration procedures often provide static or linear corrections, with the drawback that higher-order skews or event-to-event corrections are not addressed. Novel research demonstrated significant improvements in the reachable timing resolutions when combining conventional calibration approaches with machine learning, with the disadvantage of extensive calibration times infeasible for a clinical application. In this work, we made the first steps towards an in-system application and analyzed the effects of varying data sparsity on a machine learning timing calibration, aiming to accelerate the calibration time. Furthermore, we demonstrated the versatility of our calibration concept by applying the procedure for the first time to analog readout technology.Approach. We modified experimentally acquired calibration data used for training regarding their statistical and spatial sparsity, mimicking reduced measurement time and variability of the training data. Trained models were tested on unseen test data, characterized by fine spatial sampling and rich statistics. In total, 80 decision tree models with the same hyperparameter settings, were trained and holistically evaluated regarding data scientific, physics-based, and PET-based quality criteria. Main results.The calibration procedure can be heavily reduced from several days to some minutes without sacrificing quality and still significantly improving the timing resolution from (304 ± 5) ps to (216 ± 1) ps compared to conventionally used analytical calibration methods. Significance. This work serves as the first step in making the developed machine learning-based calibration suitable for an in-system application to profit from the method’s capabilities on the system level. Furthermore, this work demonstrates the functionality of the methodology on detectors using analog readout technology. The proposed holistic evaluation criteria here serve as a guideline for future evaluations of machine learning-based calibration approaches.

show abstract

Training Data Influence Analysis and Estimation: A Survey

Cited by 3 publications

References 0 publications

Machine Learning to Predict Enzyme–Substrate Interactions in Elucidation of Synthesis Pathways: A Review

Machine Learning to Predict Enzyme–Substrate Interactions in Elucidation of Synthesis Pathways: A Review

Classification of Kinematic Data Using Explainable Artificial Intelligence (XAI) for Smart Motion

Holistic evaluation of a machine learning-based timing calibration for PET detectors under varying data sparsity

Contact Info

Product

Resources

About