Interpretable Representation Research Articles

Decision Tree Regressions Estimate Liquid Holdup in Two-Phase Gas/Liquid Flows

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 203448, “Decision-Tree Regressions for Estimating Liquid Holdup in Two-Phase Gas/Liquid Flows,” by Meshal Almashan, SPE, Yoshiaki Narusue, and Hiroyuki Morikawa, University of Tokyo, prepared for the 2020 Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, held virtually 9–12 November. The paper has not been peer reviewed. In the authors’ study, a machine-learning predictive model—boosted decision tree regression (BDTR)—is trained, tested, and evaluated in predicting liquid holdup (HL) in multiphase flows in oil and gas wells. Results show that the proposed BDTR model outperforms the best empirical correlations and the fuzzy-logic model used in estimating HL in gas/liquid multiphase flows. Using the BDTR model with its interpretable representation, the heuristic feature importance of the input features used in building the model can be determined clearly. Introduction Machine-learning approaches in predicting HL in multiphase flows have been recently studied to improve prediction accuracy compared with existing empirical correlations. However, these approaches ignore the heuristic feature importance of the input parameters to the predicted HL values. The heuristic feature importance can help provide better insight into the issues associated with HL studies, such as the liquid-loading phenomenon. To the best of the authors’ knowledge, the present study is the first work that shows how decision-forest regression predictive models can predict HL accurately. Data Acquisition The performance and the predictive power of a machine-learning model relies greatly on the quality and completeness of the data set used in building the model. The data sets used in training and testing the predictive model are experimental and were collected from the literature (111 data points). Air/kerosene and air/water mixtures were used in obtaining the 111 experimental data points. In this study, this data set is divided into three different subsets: training, validation, and testing. The data sets consist of the properties of HL, the superficial gas velocity (Vsg), the superficial liquid velocity (Vsl), pressure, and temperature (T). The statistical measures of the data sets are shown in Table 1 of the complete paper.

An analytical framework for interpretable and generalizable single-cell data analysis.

Scaling single-cell data exploratory analysis with the rapidly growing diversity and quantity of single-cell omics datasets demands more interpretable and robust data representation that is generalizable across datasets. Here we developed a ‘linearly interpretable’ framework that combines the interpretability and transferability of linear methods with the representational power of nonlinear methods. Within this framework, we introduce a data representation and visualization method, GraphDR, and a structure discovery method, StructDR, that unifies cluster, trajectory, and surface estimation and allows their confidence set inference.

DeepMiner: Discovering Interpretable Representations for Mammogram Classification and Explanation

We propose DeepMiner, a framework to discover interpretable representations in deep neural networks and to build explanations for medical predictions. By probing convolutional neural networks (CNNs) trained to classify cancer in mammograms, we show that many individual units in the final convolutional layer of a CNN respond strongly to diseased tissue concepts specified by the Breast Imaging-Reporting and Data System (BI-RADS) lexicon. After expert annotation of the interpretable units, our proposed method is able to generate explanations for CNN mammogram classification that are consistent with ground truth radiology reports on the Digital Database for Screening Mammography. We show that DeepMiner not only enables better understanding of the nuances of CNN classification decisions but also possibly discovers new visual knowledge relevant to medical diagnosis.

Physical Representation Learning and Parameter Identification from Video Using Differentiable Physics

Representation learning for video is increasingly gaining attention in the field of computer vision. For instance, video prediction models enable activity and scene forecasting or vision-based planning and control. In this article, we investigate the combination of differentiable physics and spatial transformers in a deep action conditional video representation network. By this combination our model learns a physically interpretable latent representation and can identify physical parameters. We propose supervised and self-supervised learning methods for our architecture. In experiments, we consider simulated scenarios with pushing, sliding and colliding objects, for which we also analyze the observability of the physical properties. We demonstrate that our network can learn to encode images and identify physical properties like mass and friction from videos and action sequences. We evaluate the accuracy of our training methods, and demonstrate the ability of our method to predict future video frames from input images and actions.

Explaining Deep Reinforcement Learning Agents in the Atari Domain through a Surrogate Model

One major barrier to applications of deep Reinforcement Learning (RL) both inside and outside of games is the lack of explainability. In this paper, we describe a lightweight and effective method to derive explanations for deep RL agents, which we evaluate in the Atari domain. Our method relies on a transformation of the pixel-based input of the RL agent to a symbolic, interpretable input representation. We then train a surrogate model, which is itself interpretable, to replicate the behavior of the target, deep RL agent. Our experiments demonstrate that we can learn an effective surrogate that accurately approximates the underlying decision making of a target agent on a suite of Atari games.

Deep learning method for identifying the minimal representations and nonlinear mode decomposition of fluid flows

We propose a deep learning method to learn the minimal representations of fluid flows. It uses the deep variational autoencoder (VAE) to decouple the independent representations for fluid flows. We apply this method to several simple flows and show that the network successfully identifies the independent and interpretable representations. It shows that the proposed method can extract the physically suggestive information. We further employ the VAE network to improve the mode decomposing autoencoder framework. It decomposes the cylinder flow fields into two independent ordered states. The cylinder flow at different Reynolds numbers and time can be described as the composition of the two decomposed fields. The present results suggest that the proposed network can be used as an effective nonlinear dimensionality reduction tool for flow fields.

Heart Rate Topic Models

A key challenge in reducing the burden of cardiovascular disease is matching patients to treatments that are most appropriate for them. Different cardiac assessment tools have been developed to address this goal. Recent research has focused on heart rate motifs, i.e., short-term heart rate sequences that are over- or under-represented in long-term electrocardiogram (ECG) recordings of patients experiencing cardiovascular outcomes, which provide novel and valuable information for risk stratification. However, this approach can leverage only a small number of motifs for prediction and results in difficult to interpret models. We address these limitations by identifying latent structure in the large numbers of motifs found in long-term ECG recordings. In particular, we explore the application of topic models to heart rate time series to identify functional sets of heart rate sequences and to concisely describe patients using task-independent features for various cardiovascular outcomes. We evaluate the approach on a large collection of real-world ECG data, and investigate the performance of topic mixture features for the prediction of cardiovascular mortality. The topics provided an interpretable representation of the recordings and maintained valuable information for clinical assessment when compared with motif frequencies, even after accounting for commonly used clinical risk scores.

Probing T-cell response by sequence-based probabilistic modeling.

With the increasing ability to use high-throughput next-generation sequencing to quantify the diversity of the human T cell receptor (TCR) repertoire, the ability to use TCR sequences to infer antigen-specificity could greatly aid potential diagnostics and therapeutics. Here, we use a machine-learning approach known as Restricted Boltzmann Machine to develop a sequence-based inference approach to identify antigen-specific TCRs. Our approach combines probabilistic models of TCR sequences with clone abundance information to extract TCR sequence motifs central to an antigen-specific response. We use this model to identify patient personalized TCR motifs that respond to individual tumor and infectious disease antigens, and to accurately discriminate specific from non-specific responses. Furthermore, the hidden structure of the model results in an interpretable representation space where TCRs responding to the same antigen cluster, correctly discriminating the response of TCR to different viral epitopes. The model can be used to identify condition specific responding TCRs. We focus on the examples of TCRs reactive to candidate neoantigens and selected epitopes in experiments of stimulated TCR clone expansion.

Natural Language Mapping of Electrocardiogram Interpretations to a Standardized Ontology.

Interpretations of the electrocardiogram (ECG) are often prepared using software outside the electronic health record (EHR) and imported via an interface as a narrative note. Thus, natural language processing is required to create a computable representation of the findings. Challenges include misspellings, nonstandard abbreviations, jargon, and equivocation in diagnostic interpretations. Our objective was to develop an algorithm to reliably and efficiently extract such information and map it to the standardized ECG ontology developed jointly by the American Heart Association, the American College of Cardiology Foundation, and the Heart Rhythm Society. The algorithm was to be designed to be easily modifiable for use with EHRs and ECG reporting systems other than the ones studied. An algorithm using natural language processing techniques was developed in structured query language to extract and map quantitative and diagnostic information from ECG narrative reports to the cardiology societies' standardized ECG ontology. The algorithm was developed using a training dataset of 43,861 ECG reports and applied to a test dataset of 46,873 reports. Accuracy, precision, recall, and the F1-measure were all 100% in the test dataset for the extraction of quantitative data (e.g., PR and QTc interval, atrial and ventricular heart rate). Performances for matches in each diagnostic category in the standardized ECG ontology were all above 99% in the test dataset. The processing speed was approximately 20,000 reports per minute. We externally validated the algorithm from another institution that used a different ECG reporting system and found similar performance. The developed algorithm had high performance for creating a computable representation of ECG interpretations. Software and lookup tables are provided that can easily be modified for local customization and for use with other EHR and ECG reporting systems. This algorithm has utility for research and in clinical decision-support where incorporation of ECG findings is desired.

A mixture-of-experts deep generative model for integrated analysis of single-cell multiomics data

SummaryThe recent development of single-cell multiomics analysis has enabled simultaneous detection of multiple traits at the single-cell level, providing deeper insights into cellular phenotypes and functions in diverse tissues. However, currently, it is challenging to infer the joint representations and learn relationships among multiple modalities from complex multimodal single-cell data. Here, we present scMM, a novel deep generative model-based framework for the extraction of interpretable joint representations and crossmodal generation. scMM addresses the complexity of data by leveraging a mixture-of-experts multimodal variational autoencoder. The pseudocell generation strategy of scMM compensates for the limited interpretability of deep learning models, and the proposed approach experimentally discovered multimodal regulatory programs associated with latent dimensions. Analysis of recently produced datasets validated that scMM facilitates high-resolution clustering with rich interpretability. Furthermore, we show that crossmodal generation by scMM leads to more precise prediction and data integration compared with the state-of-the-art and conventional approaches.

Conceptual Understanding through Efficient Automated Design of Quantum Optical Experiments

One crucial question within artificial intelligence research is how this technology can be used to discover new scientific concepts and ideas. We present Theseus, an explainable AI algorithm that can contribute to science at a conceptual level. This work entails four significant contributions. (i) We introduce an interpretable representation of quantum optical experiments amenable to algorithmic use. (ii) We develop an inverse-design approach for new quantum experiments, which is orders of magnitudes faster than the best previous methods. (iii) We solve several crucial open questions in quantum optics, which is expected to advance photonic technology. Finally, and most importantly, (iv) the interpretable representation and drastic speedup produce solutions that a human scientist can interpret outright to discover new scientific concepts. We anticipate that Theseus will become an essential tool in quantum optics and photonic hardware, with potential applicability to other quantum physical disciplines.

Visual representation of proverbs in comic books and their translation: Asterix as a paradigmatic case

ABSTRACT Proverbs are found in the most varied means of communication. In relation to their use in comic books, they show a creative use that can only take place in certain circumstances: the visual representation of a literal interpretation of the message they convey. Comic books in general are an excellent medium for this, and the works under examination, the adventures of Asterix the Gaul, is the ideal subject of analysis for a study like the present, thanks to the linguistic inclination that its authors have shown. For this paper, the 38 volumes of the adventures have been surveyed for their use of paremiology with the intention of presenting and analysing cases of the use of paremias that seemed of relevance in connection to the graphic side of comic books. On the following pages, some of the most remarkable cases will be dissected in order to determine how this phenomenon works. Moreover, a cross-linguistic analysis of the use of said proverbs will be carried out, presenting the original version in French and comparing it with five other languages to assess how translators deal with the complicated task of translating idiomatic expressions and how the visual content of comic books determines this practice.

A Study of the Application of Technology Acceptance Model (TAM) to E learning among undergraduate students in India- A Structural Equation Modelling Approach

Technological developments in the field of education have paved the way for e-learning. Massive Open Online Courses (MOOC) and online certification courses have become widely popular among students especially those in higher education. Today, E-learning platforms have provided the end users the opportunity to pursue online certifications, gain knowledge and skills, and learn from experts in different fields from the comfort of their homes. Students are an important stakeholder in this knowledge transfer process. Therefore, it is important to understand their attitude towards e learning and their expectation from such platforms. This paper attempts to study the concept of Technology Acceptance Model (TAM) in e-learning among undergraduate students in India to understand their perception and behaviour towards e-learning. The main research methods used were: the questionnaire for primary data collection, journals and research articles for secondary data collection. Excel and Statistical Product and Service Solutions (SPSS) were used to process the collected data, test the hypotheses, statistical analysis, analyse the results and achieve research objectives. Microsoft Word was used for textual representation of final results and interpretations. Hypothetical model based on TAM was framed and six research variables were identified for the study: Perceived Usefulness, Capability, Perceived ease-of-use, Intention to Use, Trustworthiness, Actual Use. The research helped to understand the components of TAM, analyse and identify the relation between TAM and E learning, and revealed that undergraduate students in India have positive perception of E-learning.

Reverse engineering of a recording mix with differentiable digital signal processing.

A method to retrieve the parameters used to create a multitrack mix using only raw tracks and the stereo mixdown is presented. This method is able to model linear time-invariant effects such as gain, pan, equalisation, delay, and reverb. Nonlinear effects, such as distortion and compression, are not considered in this work. The optimization procedure used is the stochastic gradient descent with the aid of differentiable digital signal processing modules. This method allows for a fully interpretable representation of the mixing signal chain by explicitly modelling the audio effects rather than using differentiable blackbox modules. Two reverb module architectures are proposed, a "stereo reverb" model and an "individual reverb" model, and each is discussed. Objective feature measures are taken of the outputs of the two architectures when tasked with estimating a target mix and compared against a stereo gain mix baseline. A listening study is performed to measure how closely the two architectures can perceptually match a reference mix when compared to a stereo gain mix. Results show that the stereo reverb model performs best on objective measures and there is no statistically significant difference between the participants' perception of the stereo reverb model and reference mixes.

Modernisasi Penafsiran Al-Quran Dalam Tafsir Al-Azhar Karya Buya Hamka

The Discourse of this research, focuses on the modernization of the interpretation of the Qur’an from the perspective of the Tafsir al-Azhar by Buya Hamka. The purpose of the study, to reveal the impact of modernization on the realm of interpretation of the Qur’an and to determine the development of the interpretation of the Qur’an by looking at the extent to which the Tafsir al-Azhar can be categorized in the realm of modernization of the interpretation of the Qur’an. This research is in the form of qualitative research, the type of library research is assisted by data analysis in the form of descriptive analysis, and to determine the significance of Tafsir al-Azhar, the author uses the interpretation theory of the interpretation of the Qur’an Nasr Hamid Abu Zayd, in order to be able to see the representation of the text of Tafsir al-Azhar has any reference to modernization or not, and is equipped with theorical triangulation as a test of the validity of the data. The results of the study state that the existence of Tafsir al-Azhar is representation of modern interpretation because of ist efforts to contextualize the text with events that are currently happening, driven by the emergence of modernization, and modernization of the interpretation of the Qur’an is also an implementation of the involvement of interpretation with modern problems, as the interpretation of Buya Hamka in the Tafsir al-Azhar which discusses a lot of modern problems. This is an attempt by the mufasir to maintain the Qur’an which should indeed be salih likulli zaman wa makan.

TAUHID HERMENEUTICS OF FAḤRUDDIN AL-RᾹZI: A STUDY ON THE INTERPRETATION OF SURAH AL-IKHLᾹS IN THE MAFᾹTIḤ AL-GAῙB

The Qur'ān is a way of life that is always mined and studied to answer the complex and dynamic problems of life. The works of interpretation are one of the proofs of interpretation products dedicated by interpreters of the ulama to succeed in the effort to understand the Qur'ān so that it can be guided and truly become the way of life. Efforts to understand the Qur'an are carried out with a variety of approaches and methods, one of which is the hermeneutic method. The Mafātiḥ al-Gaīb interpretation by Faḥruddin Al-Rāzi is a representation of interpretation that applies hermeneutic elements. This paper aims to explore the elements of hermeneutics in Mafātiḥ al-Gaīb's interpretation by Faḥruddin Al-Rāzi by taking a sample of Al-Ikhlās's letter. The research method used in this study is a qualitative descriptive method, the authors attempt to examine and describe the data that exists. The findings in this study indicate that the hermeneutic elements used by Al-Rāzi in the interpretation of Al-Ikhlās include munāsabah, asbāb an-nuzūl, and esoteric meanings. The three components are used to find out the meaning of an objective and comprehensive text from Surah Al-Ikhlās. So that it concludes that the essence of God referred to in Surah Al-Ikhlās includes His żat, ṣifat, and af’al.

Multivariate times series classification through an interpretable representation

Multivariate time series classification is a machine learning task with increasing importance due to the proliferation of information sources in different domains (economy, health, energy, crops, etc.). Univariate methods lack the ability to capture the relationships between the different variables that compose a multivariate time series and therefore cannot be directly extrapolated to multivariate environments. Despite the good performance and competitive results of the multivariate proposals published to date, they are hard to interpret due to their high complexity. In this paper, we propose a multivariate time series classification method based on an alternative representation of the time series, composed of a set of 41 descriptive time series features, in order to improve the interpretability of time series and results obtained. Our proposal uses traditional classifiers over the extracted features to look for relationships between the different variables that form a multivariate time series. We have selected four state-of-the-art algorithms as base classifiers to evaluate our method. We have tested our proposal on the complete University of East Anglia repository, obtaining highly interpretable results capable of explaining the relationships between the features that compose the time series and achieving performance results statistically indistinguishable from the best algorithms of the state-of-the-art.

Extending Multi-Sense Word Embedding to Phrases and Sentences for Unsupervised Semantic Applications

Most unsupervised NLP models represent each word with a single point or single region in semantic space, while the existing multi-sense word embeddings cannot represent longer word sequences like phrases or sentences. We propose a novel embedding method for a text sequence (a phrase or a sentence) where each sequence is represented by a distinct set of multi-mode codebook embeddings to capture different semantic facets of its meaning. The codebook embeddings can be viewed as the cluster centers which summarize the distribution of possibly co-occurring words in a pre-trained word embedding space. We introduce an end-to-end trainable neural model that directly predicts the set of cluster centers from the input text sequence during test time. Our experiments show that the per-sentence codebook embeddings significantly improve the performances in unsupervised sentence similarity and extractive summarization benchmarks. In phrase similarity experiments, we discover that the multi-facet embeddings provide an interpretable semantic representation but do not outperform the single-facet baseline.

Learning Game-Theoretic Models of Multiagent Trajectories Using Implicit Layers

For prediction of interacting agents' trajectories, we propose an end-to-end trainable architecture that hybridizes neural nets with game-theoretic reasoning, has interpretable intermediate representations, and transfers to downstream decision making. It uses a net that reveals preferences from the agents' past joint trajectory, and a differentiable implicit layer that maps these preferences to local Nash equilibria, forming the modes of the predicted future trajectory. Additionally, it learns an equilibrium refinement concept. For tractability, we introduce a new class of continuous potential games and an equilibrium-separating partition of the action space. We provide theoretical results for explicit gradients and soundness. In experiments, we evaluate our approach on two real-world data sets, where we predict highway drivers' merging trajectories, and on a simple decision-making transfer task.

TripleTree: A Versatile Interpretable Representation of Black Box Agents and their Environments

In explainable artificial intelligence, there is increasing interest in understanding the behaviour of autonomous agents to build trust and validate performance. Modern agent architectures, such as those trained by deep reinforcement learning, are currently so lacking in interpretable structure as to effectively be black boxes, but insights may still be gained from an external, behaviourist perspective. Inspired by conceptual spaces theory, we suggest that a versatile first step towards general understanding is to discretise the state space into convex regions, jointly capturing similarities over the agent's action, value function and temporal dynamics within a dataset of observations. We create such a representation using a novel variant of the CART decision tree algorithm, and demonstrate how it facilitates practical understanding of black box agents through prediction, visualisation and rule-based explanation.