AID: Active Distillation Machine to Leverage Pre-Trained Black-Box Models in Private Data Settings

This paper presents an active distillation method for a local institution (e.g., hospital) to find the best queries within its given budget to distill an on-server black-box model’s predictive knowledge into a local surrogate with transparent parameterization. This allows local institutions to understand better the predictive reasoning of the black-box model in its own local context or to further customize the distilled knowledge with its private dataset that cannot be centralized and fed into the server model. The proposed method thus addresses several challenges of deploying machine learning (ML) in many industrial settings (e.g., healthcare analytics) with strong proprietary constraints. These include: (1) the opaqueness of the server model’s architecture which prevents local users from understanding its predictive reasoning in their local data contexts; (2) the increasing cost and risk of uploading local data on the cloud for analysis; and (3) the need to customize the server model with private onsite data. We evaluated the proposed method on both benchmark and real-world healthcare data where significant improvements over existing local distillation methods were observed. A theoretical analysis of the proposed method is also presented.

[1]  Scott Lundberg,et al.  A Unified Approach to Interpreting Model Predictions , 2017, NIPS.

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

[3]  Franco Turini,et al.  A Survey of Methods for Explaining Black Box Models , 2018, ACM Comput. Surv..

[4]  Kian Hsiang Low,et al.  Collective Model Fusion for Multiple Black-Box Experts , 2019, ICML.

[5]  Motoaki Kawanabe,et al.  How to Explain Individual Classification Decisions , 2009, J. Mach. Learn. Res..

[6]  Le Song,et al.  GRAM: Graph-based Attention Model for Healthcare Representation Learning , 2016, KDD.

[7]  Geoffrey E. Hinton,et al.  Distilling the Knowledge in a Neural Network , 2015, ArXiv.

[8]  Peter Szolovits,et al.  MIMIC-III, a freely accessible critical care database , 2016, Scientific Data.

[9]  Andrew Zisserman,et al.  Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps , 2013, ICLR.

[10]  Patrick Jaillet,et al.  Learning Task-Agnostic Embedding of Multiple Black-Box Experts for Multi-Task Model Fusion , 2020, ICML.

[11]  Alexander Binder,et al.  On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation , 2015, PloS one.

[12]  Yoshua Bengio,et al.  BinaryNet: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1 , 2016, ArXiv.

[13]  Eric J Topol,et al.  High-performance medicine: the convergence of human and artificial intelligence , 2019, Nature Medicine.

[14]  Vineeth N. Balasubramanian,et al.  Deep Model Compression: Distilling Knowledge from Noisy Teachers , 2016, ArXiv.

[15]  Jimeng Sun,et al.  RAIM: Recurrent Attentive and Intensive Model of Multimodal Patient Monitoring Data , 2018, KDD.

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

[17]  Le Song,et al.  Learning to Explain: An Information-Theoretic Perspective on Model Interpretation , 2018, ICML.

[18]  Rich Caruana,et al.  Do Deep Nets Really Need to be Deep? , 2013, NIPS.

[19]  Junmo Kim,et al.  A Gift from Knowledge Distillation: Fast Optimization, Network Minimization and Transfer Learning , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[20]  Kaiming He,et al.  Data Distillation: Towards Omni-Supervised Learning , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[21]  Thomas Brox,et al.  Striving for Simplicity: The All Convolutional Net , 2014, ICLR.

[22]  David Cohn,et al.  Active Learning , 2010, Encyclopedia of Machine Learning.

[23]  Jimeng Sun,et al.  RETAIN: An Interpretable Predictive Model for Healthcare using Reverse Time Attention Mechanism , 2016, NIPS.

[24]  Avanti Shrikumar,et al.  Learning Important Features Through Propagating Activation Differences , 2017, ICML.

[25]  Mohan S. Kankanhalli,et al.  Active Learning Is Planning: Nonmyopic ε-Bayes-Optimal Active Learning of Gaussian Processes , 2014, ECML/PKDD.

[26]  Bernhard Schölkopf,et al.  Unifying distillation and privileged information , 2015, ICLR.