ProtoSteer: Steering Deep Sequence Model with Prototypes

Recently we have witnessed growing adoption of deep sequence models (e.g. LSTMs) in many application domains, including predictive health care, natural language processing, and log analysis. However, the intricate working mechanism of these models confines their accessibility to the domain experts. Their black-box nature also makes it a challenging task to incorporate domain-specific knowledge of the experts into the model. In ProtoSteer (Prototype Steering), we tackle the challenge of directly involving the domain experts to steer a deep sequence model without relying on model developers as intermediaries. Our approach originates in case-based reasoning, which imitates the common human problem-solving process of consulting past experiences to solve new problems. We utilize ProSeNet (Prototype Sequence Network), which learns a small set of exemplar cases (i.e., prototypes) from historical data. In ProtoSteer they serve both as an efficient visual summary of the original data and explanations of model decisions. With ProtoSteer the domain experts can inspect, critique, and revise the prototypes interactively. The system then incorporates user-specified prototypes and incrementally updates the model. We conduct extensive case studies and expert interviews in application domains including sentiment analysis on texts and predictive diagnostics based on vehicle fault logs. The results demonstrate that involvements of domain users can help obtain more interpretable models with concise prototypes while retaining similar accuracy.

[1]  Barbara Hammer,et al.  Parametric nonlinear dimensionality reduction using kernel t-SNE , 2015, Neurocomputing.

[2]  Yuanzhe Chen,et al.  Sequence Synopsis: Optimize Visual Summary of Temporal Event Data , 2018, IEEE Transactions on Visualization and Computer Graphics.

[3]  David Gotz,et al.  Data-driven exploration of care plans for patients , 2013, CHI Extended Abstracts.

[4]  Jerry Alan Fails,et al.  Interactive machine learning , 2003, IUI '03.

[5]  Cynthia Rudin,et al.  Deep Learning for Case-based Reasoning through Prototypes: A Neural Network that Explains its Predictions , 2017, AAAI.

[6]  Maya Cakmak,et al.  Power to the People: The Role of Humans in Interactive Machine Learning , 2014, AI Mag..

[7]  Geoffrey E. Hinton,et al.  Visualizing Data using t-SNE , 2008 .

[8]  Zhimin Li,et al.  NLIZE: A Perturbation-Driven Visual Interrogation Tool for Analyzing and Interpreting Natural Language Inference Models , 2019, IEEE Transactions on Visualization and Computer Graphics.

[9]  Alexander M. Rush,et al.  Seq2seq-Vis: A Visual Debugging Tool for Sequence-to-Sequence Models , 2018, IEEE Transactions on Visualization and Computer Graphics.

[10]  Zachary Chase Lipton.  A Critical Review of Recurrent Neural Networks for Sequence Learning , 2015, ArXiv.

[11]  Bin Yu,et al.  Beyond Word Importance: Contextual Decomposition to Extract Interactions from LSTMs , 2018, ICLR.

[12]  D. Botstein,et al.  Cluster analysis and display of genome-wide expression patterns. , 1998, Proceedings of the National Academy of Sciences of the United States of America.

[13]  David Gotz,et al.  Exploring Flow, Factors, and Outcomes of Temporal Event Sequences with the Outflow Visualization , 2012, IEEE Transactions on Visualization and Computer Graphics.

[14]  Kevin Knight,et al.  Artificial intelligence (2. ed.) , 1991 .

[15]  Gang Wang,et al.  Unsupervised Clickstream Clustering for User Behavior Analysis , 2016, CHI.

[16]  Carlos Guestrin,et al.  Anchors: High-Precision Model-Agnostic Explanations , 2018, AAAI.

[17]  Carlos Guestrin,et al.  "Why Should I Trust You?": Explaining the Predictions of Any Classifier , 2016, ArXiv.

[18]  Fei Wang,et al.  Frequence: interactive mining and visualization of temporal frequent event sequences , 2014, IUI.

[19]  Jarke J. van Wijk,et al.  BaobabView: Interactive construction and analysis of decision trees , 2011, 2011 IEEE Conference on Visual Analytics Science and Technology (VAST).

[20]  J. Stasko,et al.  Focus+context display and navigation techniques for enhancing radial, space-filling hierarchy visualizations , 2000, IEEE Symposium on Information Visualization 2000. INFOVIS 2000. Proceedings.

[21]  David Gotz,et al.  DecisionFlow: Visual Analytics for High-Dimensional Temporal Event Sequence Data , 2014, IEEE Transactions on Visualization and Computer Graphics.

[22]  Xinlei Chen,et al.  Visualizing and Understanding Neural Models in NLP , 2015, NAACL.

[23]  Arthur Szlam,et al.  Automatic Rule Extraction from Long Short Term Memory Networks , 2016, ICLR.

[24]  Hans-Peter Kriegel,et al.  Visual classification: an interactive approach to decision tree construction , 1999, KDD '99.

[25]  Hongyuan Zha,et al.  Visual Progression Analysis of Event Sequence Data , 2019, IEEE Transactions on Visualization and Computer Graphics.

[26]  Paul N. Bennett,et al.  Guidelines for Human-AI Interaction , 2019, CHI.

[27]  Alexander M. Rush,et al.  LSTMVis: A Tool for Visual Analysis of Hidden State Dynamics in Recurrent Neural Networks , 2016, IEEE Transactions on Visualization and Computer Graphics.

[28]  Bolin Ding,et al.  Trust, but Verify: Optimistic Visualizations of Approximate Queries for Exploring Big Data , 2017, CHI.

[29]  Minsuk Kahng,et al.  ActiVis: Visual Exploration of Industry-Scale Deep Neural Network Models , 2017, IEEE Transactions on Visualization and Computer Graphics.

[30]  Jimeng Sun,et al.  RetainVis: Visual Analytics with Interpretable and Interactive Recurrent Neural Networks on Electronic Medical Records , 2018, IEEE Transactions on Visualization and Computer Graphics.

[31]  Ke Xu,et al.  EventThread: Visual Summarization and Stage Analysis of Event Sequence Data , 2018, IEEE Transactions on Visualization and Computer Graphics.

[32]  Helen Yannakoudakis,et al.  Automatic Text Scoring Using Neural Networks , 2016, ACL.

[33]  Fei-Fei Li,et al.  Visualizing and Understanding Recurrent Networks , 2015, ArXiv.

[34]  Janet L. Kolodner,et al.  An introduction to case-based reasoning , 1992, Artificial Intelligence Review.

[35]  Minsuk Kahng,et al.  Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontiers , 2018, IEEE Transactions on Visualization and Computer Graphics.

[36]  Per Ola Kristensson,et al.  A Review of User Interface Design for Interactive Machine Learning , 2018, ACM Trans. Interact. Intell. Syst..

[37]  Bum Chul Kwon,et al.  Peekquence : Visual Analytics for Event Sequence Data , 2016 .

[38]  Dylan Cashman,et al.  RNNbow: Visualizing Learning Via Backpropagation Gradients in RNNs , 2018, IEEE Computer Graphics and Applications.

[39]  Desney S. Tan,et al.  Interactive optimization for steering machine classification , 2010, CHI.

[40]  Desney S. Tan,et al.  Effective End-User Interaction with Machine Learning , 2011, AAAI.

[41]  Minsuk Kahng,et al.  TimeStitch: Interactive multi-focus cohort discovery and comparison , 2015, 2015 IEEE Conference on Visual Analytics Science and Technology (VAST).

[42]  Huamin Qu,et al.  Interpretable and Steerable Sequence Learning via Prototypes , 2019, KDD.

[43]  Min Chen,et al.  VIS4ML: An Ontology for Visual Analytics Assisted Machine Learning , 2019, IEEE Transactions on Visualization and Computer Graphics.

[44]  Ben Shneiderman,et al.  Coping with Volume and Variety in Temporal Event Sequences: Strategies for Sharpening Analytic Focus , 2017, IEEE Transactions on Visualization and Computer Graphics.

[45]  Anastasia Bezerianos,et al.  Evaluation of Interactive Machine Learning Systems , 2018, Human and Machine Learning.

[46]  Desney S. Tan,et al.  EnsembleMatrix: interactive visualization to support machine learning with multiple classifiers , 2009, CHI.

[47]  Mira Dontcheva,et al.  MatrixWave: Visual Comparison of Event Sequence Data , 2015, CHI.

[48]  Yang Wang,et al.  Patterns and Sequences: Interactive Exploration of Clickstreams to Understand Common Visitor Paths , 2017, IEEE Transactions on Visualization and Computer Graphics.