Current Machine Learning Systems Research Articles

Machine Learning Models for Parkinson Disease: Systematic Review.

With the increasing availability of data, computing resources, and easier-to-use software libraries, machine learning (ML) is increasingly used in disease detection and prediction, including for Parkinson disease (PD). Despite the large number of studies published every year, very few ML systems have been adopted for real-world use. In particular, a lack of external validity may result in poor performance of these systems in clinical practice. Additional methodological issues in ML design and reporting can also hinder clinical adoption, even for applications that would benefit from such data-driven systems. To sample the current ML practices in PD applications, we conducted a systematic review of studies published in 2020 and 2021 that used ML models to diagnose PD or track PD progression. We conducted a systematic literature review in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines in PubMed between January 2020 and April 2021, using the following exact string: "Parkinson's" AND ("ML" OR "prediction" OR "classification" OR "detection" or "artificial intelligence" OR "AI"). The search resulted in 1085 publications. After a search query and review, we found 113 publications that used ML for the classification or regression-based prediction of PD or PD-related symptoms. Only 65.5% (74/113) of studies used a holdout test set to avoid potentially inflated accuracies, and approximately half (25/46, 54%) of the studies without a holdout test set did not state this as a potential concern. Surprisingly, 38.9% (44/113) of studies did not report on how or if models were tuned, and an additional 27.4% (31/113) used ad hoc model tuning, which is generally frowned upon in ML model optimization. Only 15% (17/113) of studies performed direct comparisons of results with other models, severely limiting the interpretation of results. This review highlights the notable limitations of current ML systems and techniques that may contribute to a gap between reported performance in research and the real-life applicability of ML models aiming to detect and predict diseases such as PD.

Semi-supervised class incremental broad network for continuous diagnosis of rotating machinery faults with limited labeled samples

Machine learning has demonstrated remarkable success in the field of intelligent fault diagnosis. In current machine learning systems, models tend to be fixed after training, which makes them can only generalize to classes that appear in the training set, and cannot continuously learn newly emerging classes. However, in real industrial scenarios, industrial data is constantly growing, which requires models to be able to acquire new knowledge while retaining existing knowledge. Furthermore, a substantial proportion of the acquired samples in industry are unlabeled and only few samples are labeled. To tackle the above two challenges, this article proposes a novel semi-supervised class incremental broad network (SSCIBN) for incremental intelligent diagnosis of rotating machinery faults under limited labeled samples. Specifically, a semi-supervised graph embedding loss function is designed. On the one hand, this loss function can learn the structural information of a large amount of unlabeled data, which overcomes the limitation that the scarcity of labeled samples leads to poor model performance. On the other hand, local structure information within old and new classes is considered in the continuous learning process to fully learn the unique manifold structure information of different classes, which in turn enhances the discriminative performance between old and new classes. Furthermore, a novel semi-supervised class incremental learning mechanism is proposed, which does not need to utilize the old class data during the incremental learning process, but can effectively retain the old class knowledge. The effectiveness of the proposed method is evaluated through multiple mechanical failure increment cases. Experimental analysis demonstrates that the designed SSCIBN has significant advantages in the class incremental application with the lack of labeled samples.

A survey on machine learning in array databases

This paper provides an in-depth survey on the integration of machine learning and array databases. First,machine learning support in modern database management systems is introduced. From straightforward implementations of linear algebra operations in SQL to machine learning capabilities of specialized database managers designed to process specific types of data, a number of different approaches are overviewed. Then, the paper covers the database features already implemented in current machine learning systems. Features such as rewriting, compression, and caching allow users to implement more efficient machine learning applications. The underlying linear algebra computations in some of the most used machine learning algorithms are studied in order to determine which linear algebra operations should be efficiently implemented by array databases. An exhaustive overview of array data and relevant array database managers is also provided. Those database features that have been proven of special importance for efficient execution of machine learning algorithms are analyzed in detail for each relevant array database management system. Finally, current state of array databases capabilities for machine learning implementation is shown through two example implementations in Rasdaman and SciDB.

Does Machine Understanding Require Consciousness?

This article addresses the question of whether machine understanding requires consciousness. Some researchers in the field of machine understanding have argued that it is not necessary for computers to be conscious as long as they can match or exceed human performance in certain tasks. But despite the remarkable recent success of machine learning systems in areas such as natural language processing and image classification, important questions remain about their limited performance and about whether their cognitive abilities entail genuine understanding or are the product of spurious correlations. Here I draw a distinction between natural, artificial, and machine understanding. I analyse some concrete examples of natural understanding and show that although it shares properties with the artificial understanding implemented in current machine learning systems it also has some essential differences, the main one being that natural understanding in humans entails consciousness. Moreover, evidence from psychology and neurobiology suggests that it is this capacity for consciousness that, in part at least, explains for the superior performance of humans in some cognitive tasks and may also account for the authenticity of semantic processing that seems to be the hallmark of natural understanding. I propose a hypothesis that might help to explain why consciousness is important to understanding. In closing, I suggest that progress toward implementing human-like understanding in machines—machine understanding—may benefit from a naturalistic approach in which natural processes are modelled as closely as possible in mechanical substrates.

Tensor relational algebra for distributed machine learning system design

We consider the question: what is the abstraction that should be implemented by the computational engine of a machine learning system? Current machine learning systems typically push whole tensors through a series of compute kernels such as matrix multiplications or activation functions, where each kernel runs on an AI accelerator (ASIC) such as a GPU. This implementation abstraction provides little built-in support for ML systems to scale past a single machine, or for handling large models with matrices or tensors that do not easily fit into the RAM of an ASIC. In this paper, we present an alternative implementation abstraction called the tensor relational algebra (TRA). The TRA is a set-based algebra based on the relational algebra. Expressions in the TRA operate over binary tensor relations, where keys are multi-dimensional arrays and values are tensors. The TRA is easily executed with high efficiency in a parallel or distributed environment, and amenable to automatic optimization. Our empirical study shows that the optimized TRA-based back-end can significantly outperform alternatives for running ML workflows in distributed clusters.

Finding faults: A scoping study of fault diagnostics for Industrial Cyber–Physical Systems

Context:As Industrial Cyber–Physical Systems (ICPS) become more connected and widely-distributed, often operating in safety-critical environments, we require innovative approaches to detect and diagnose the faults that occur in them. Objective:We profile fault identification and diagnosis techniques employed in the aerospace, automotive, and industrial control domains. Each of these sectors has adopted particular methods to meet their differing diagnostic needs. By examining both theoretical presentations as well as case studies from production environments, we present a profile of the current approaches being employed and identify gaps. Methodology:A scoping study was used to identify and compare fault detection and diagnosis methodologies that are presented in the current literature. We created categories for the different diagnostic approaches via a pilot study and present an analysis of the trends that emerged. We then compared the maturity of these approaches by adapting and using the NASA Technology Readiness Level (TRL) scale. Results:Fault identification and analysis studies from 127 papers published from 2004 to 2019 reveal a wide diversity of promising techniques, both emerging and in-use. These range from traditional Physics-based Models to Data-Driven Artificial Intelligence (AI) and Knowledge-Based approaches. Hybrid techniques that blend aspects of these three broad categories were also encountered. Predictive diagnostics or prognostics featured prominently across all sectors, along with discussions of techniques including Fault trees, Petri nets and Markov approaches. We also profile some of the techniques that have reached the highest Technology Readiness Levels, showing how those methods are being applied in real-world environments beyond the laboratory. Conclusions:Our results suggest that the continuing wide use of both Model-Based and Data-Driven AI techniques across all domains, especially when they are used together in hybrid configuration, reflects the complexity of the current ICPS application space. While creating sufficiently-complete models is labor intensive, Model-free AI techniques were evidenced as a viable way of addressing aspects of this challenge, demonstrating the increasing sophistication of current machine learning systems. Connecting ICPS together to share sufficient telemetry to diagnose and manage faults is difficult when the physical environment places demands on ICPS. Despite these challenges, the most mature papers present robust fault diagnosis and analysis techniques which have moved beyond the laboratory and are proving valuable in real-world environments.

Never-ending learning

Whereas people learn many different types of knowledge from diverse experiences over many years, and become better learners over time, most current machine learning systems are much more narrow, learning just a single function or data model based on statistical analysis of a single data set. We suggest that people learn better than computers precisely because of this difference, and we suggest a key direction for machine learning research is to develop software architectures that enable intelligent agents to also learn many types of knowledge, continuously over many years, and to become better learners over time. In this paper we define more precisely this never-ending learning paradigm for machine learning, and we present one case study: the Never-Ending Language Learner (NELL), which achieves a number of the desired properties of a never-ending learner. NELL has been learning to read the Web 24hrs/day since January 2010, and so far has acquired a knowledge base with 120mn diverse, confidence-weighted beliefs (e.g., servedWith(tea,biscuits) ), while learning thousands of interrelated functions that continually improve its reading competence over time. NELL has also learned to reason over its knowledge base to infer new beliefs it has not yet read from those it has, and NELL is inventing new relational predicates to extend the ontology it uses to represent beliefs. We describe the design of NELL, experimental results illustrating its behavior, and discuss both its successes and shortcomings as a case study in never-ending learning. NELL can be tracked online at http://rtw.ml.cmu.edu, and followed on Twitter at @CMUNELL.

On invariance and selectivity in representation learning

AbstractWe discuss data representation which can be learned automaticallyfrom data, are invariant to transformations, and at the same time selec-tive, in the sense that two points have the same representation only if theyare one the transformation of the other. The mathematical results heresharpen some of the key claims of i-theory { a recent theory of feedforwardprocessing in sensory cortex. [3, 4, 5].Keywords:Invariance, Machine LearningThe paper is submitted to Information and Inference Journal. 1 Introduction This paper considers the problem of learning good data representation whichcan lower the need of labeled data (sample complexity) in machine learning(ML). Indeed, while current ML systems have achieved impressive results in avariety of tasks, an obvious bottleneck appears to be the huge amount of labeleddata needed. This paper builds on the idea that data representation, which arelearned in an unsupervised manner, can be key to solve the problem. Classicalstatistical learning theory focuses on supervised learning and postulates that asuitable hypothesis space is given. In turn, under very general conditions, thelatter can be seen to be equivalent to a data representation. In other words,data representation and how to select and learn it, is classically not consideredto be part of the learning problem, but rather as a prior information. In practicead hoc solutions are often empirically found for each problem.The study in this paper is a step towards developing a theory of learningdata representation. Our starting point is the intuition that, since many learning

Never-Ending Learning

Whereas people learn many different types of knowledge from diverse experiences over many years, most current machine learning systems acquire just a single function or data model from just a single data set. We propose a never-ending learning paradigm for machine learning, to better reflect the more ambitious and encompassing type of learning performed by humans. As a case study, we describe the Never-Ending Language Learner (NELL), which achieves some of the desired properties of a never-ending learner, and we discuss lessons learned. NELL has been learning to read the web 24 hours/day since January 2010, and so far has acquired a knowledge base with over 80 million confidence-weighted beliefs (e.g., servedWith(tea, biscuits)). NELL has also learned millions of features and parameters that enable it to read these beliefs from the web. Additionally, it has learned to reason over these beliefs to infer new beliefs, and is able to extend its ontology by synthesizing new relational predicates. NELL can be tracked online at http://rtw.ml.cmu.edu, and followed on Twitter at @CMUNELL.