Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userPrIU: A Provenance-Based Approach for Incrementally Updating Regression Models
60
0
0.0
(
0
)
Added by
Yinjun Wu
Publication date
2020
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
The ubiquitous use of machine learning algorithms brings new challenges to traditional database problems such as incremental view update. Much effort is being put in better understanding and debugging machine learning models, as well as in identifying and repairing errors in training datasets. Our focus is on how to assist these activities when they have to retrain the machine learning model after removing problematic training samples in cleaning or selecting different subsets of training data for interpretability. This paper presents an efficient provenance-based approach, PrIU, and its optimized version, PrIU-opt, for incrementally updating model parameters without sacrificing prediction accuracy. We prove the correctness and convergence of the incrementally updated model parameters, and validate it experimentally. Experimental results show that up to two orders of magnitude speed-ups can be achieved by PrIU-opt compared to simply retraining the model from scratch, yet obtaining highly similar models.
rate research
Read More
For data-centric systems, provenance tracking is particularly important when the system is open and decentralised, such as the Web of Linked Data. In this paper, a concise but expressive calculus which models data updates is presented. The calculus is used to provide an operational semantics for a system where data and updates interact concurrently. The operational semantics of the calculus also tracks the provenance of data with respect to updates. This provides a new formal semantics extending provenance diagrams which takes into account the execution of processes in a concurrent setting. Moreover, a sound and complete model for the calculus based on ideals of series-parallel DAGs is provided. The notion of provenance introduced can be used as a subjective indicator of the quality of data in concurrent interacting systems.
Provenance is information about the origin, derivation, ownership, or history of an object. It has recently been studied extensively in scientific databases and other settings due to its importance in helping scientists judge data validity, quality and integrity. However, most models of provenance have been stated as ad hoc definitions motivated by informal concepts such as comes from, influences, produces, or depends on. These models lack clear formalizations describing in what sense the definitions capture these intuitive concepts. This makes it difficult to compare approaches, evaluate their effectiveness, or argue about their validity. We introduce provenance traces, a general form of provenance for the nested relational calculus (NRC), a core database query language. Provenance traces can be thought of as concrete data structures representing the operational semantics derivation of a computation; they are related to the traces that have been used in self-adjusting computation, but differ in important respects. We define a tracing operational semantics for NRC queries that produces both an ordinary result and a trace of the execution. We show that three pre-existing forms of provenance for the NRC can be extracted from provenance traces. Moreover, traces satisfy two semantic guarantees: consistency, meaning that the traces describe what actually happened during execution, and fidelity, meaning that the traces explain how the expression would behave if the input were changed. These guarantees are much stronger than those contemplated for previous approaches to provenance; thus, provenance traces provide a general semantic foundation for comparing and unifying models of provenance in databases.
Deep learning owes much of its success to the astonishing expressiveness of neural networks. However, this comes at the cost of complex, black-boxed models that extrapolate poorly beyond the domain of the training dataset, conflicting with goals of finding analytic expressions to describe science, engineering and real world data. Under the hypothesis that the hierarchical modularity of such laws can be captured by training a neural network, we introduce OccamNet, a neural network model that finds interpretable, compact, and sparse solutions for fitting data, `{a} la Occams razor. Our model defines a probability distribution over a non-differentiable function space. We introduce a two-step optimization method that samples functions and updates the weights with backpropagation based on cross-entropy matching in an evolutionary strategy: we train by biasing the probability mass toward better fitting solutions. OccamNet is able to fit a variety of symbolic laws including simple analytic functions, recursive programs, implicit functions, simple image classification, and can outperform noticeably state-of-the-art symbolic regression methods on real world regression datasets. Our method requires minimal memory footprint, does not require AI accelerators for efficient training, fits complicated functions in minutes of training on a single CPU, and demonstrates significant performance gains when scaled on a GPU. Our implementation, demonstrations and instructions for reproducing the experiments are available at https://github.com/druidowm/OccamNet_Public.
Learning accurate models of the physical world is required for a lot of robotic manipulation tasks. However, during manipulation, robots are expected to interact with unknown workpieces so that building predictive models which can generalize over a number of these objects is highly desirable. In this paper, we study the problem of designing deep learning agents which can generalize their models of the physical world by building context-aware learning models. The purpose of these agents is to quickly adapt and/or generalize their notion of physics of interaction in the real world based on certain features about the interacting objects that provide different contexts to the predictive models. With this motivation, we present context-aware zero shot learning (CAZSL, pronounced as casual) models, an approach utilizing a Siamese network architecture, embedding space masking and regularization based on context variables which allows us to learn a model that can generalize to different parameters or features of the interacting objects. We test our proposed learning algorithm on the recently released Omnipush datatset that allows testing of meta-learning capabilities using low-dimensional data. Codes for CAZSL are available at https://www.merl.com/research/license/CAZSL.
Heterogeneity is an important feature of modern data sets and a central task is to extract information from large-scale and heterogeneous data. In this paper, we consider multiple high-dimensional linear models and adopt the definition of maximin effect (Meinshausen, B{u}hlmann, AoS, 43(4), 1801--1830) to summarize the information contained in this heterogeneous model. We define the maximin effect for a targeted population whose covariate distribution is possibly different from that of the observed data. We further introduce a ridge-type maximin effect to simultaneously account for reward optimality and statistical stability. To identify the high-dimensional maximin effect, we estimate the regression covariance matrix by a debiased estimator and use it to construct the aggregation weights for the maximin effect. A main challenge for statistical inference is that the estimated weights might have a mixture distribution and the resulted maximin effect estimator is not necessarily asymptotic normal. To address this, we devise a novel sampling approach to construct the confidence interval for any linear contrast of high-dimensional maximin effects. The coverage and precision properties of the proposed confidence interval are studied. The proposed method is demonstrated over simulations and a genetic data set on yeast colony growth under different environments.
suggested questions
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
