- Location: NIPS 2017 Workshop, Long Beach, CA, USA
- Room: Seaside Ballroom
- Date: 9 Dec, 2017 (Saturday)
- Accepted Papers: List of accepted papers is now available here!
- Schedule: Final schedule is now available here!
- Contact: For any questions, please email nips17teaching AT gmail.com
- Updates (after the workshop):
This workshop focuses on "machine teaching", the inverse problem of machine learning, in which the goal is to find an optimal training set given a machine learning algorithm and a target model. The study of machine teaching began in the early 1990s, primarily coming out of computational learning theory. Recently, there has been a surge of interest in machine teaching as several different communities within machine learning have found connections to this problem; these connections have included the following:
- machine teaching has close connections to newly introduced models of interaction in machine learning community, such as curriculum learning, self-paced learning, and knowledge distillation. [Hinton et al. 2015; Bengio et al. 2009]
- there are strong theoretical connections between the Teaching-dimension (the sample complexity of teaching) and the VC-dimension (the sample complexity of learning from randomly chosen examples). [Doliwa et al. 2014]
- machine teaching problem formulation has been recently studied in the context of diverse applications including personalized educational systems, cyber-security problems, robotics, program synthesis, human-in-the-loop systems, and crowdsourcing. [Jha et al. 2016; Zhu 2015; Mei & Zhu 2015; Ba & Caruana 2014; Patil et al. 2014; Singla et al. 2014; Cakmak & Thomaz 2014]
In this workshop, we draw attention to machine teaching by emphasizing how the area of machine teaching interacts with emerging research trends and application domains relevant to the NIPS community. The goal of this workshop is to foster these ideas by bringing together researchers with expertise/interest in the inter-related areas of machine teaching, interactive machine learning, robotics, cyber-security problems, generative adversarial networks, educational technologies, and cognitive science.
Topics of InterestTopics of interests in the workshop include (but are not limited to):
- using tools from information theory to develop better mathematical models of teaching;
- characterizing the complexity of teaching when a teacher has limited power, or incomplete knowledge of student's model, or a mismatch in feature representations;
- algorithms for adaptive teaching by interactively inferring the learner's state;
- new notions of Teaching-dimension for generic teaching settings.
- the information complexity of teaching and query complexity;
- machine teaching vs. curriculum learning and other models of interactive machine learning;
- teaching reinforcement learning agents
- using the machine teaching formulation to enable more rigorous and generalizable approaches for developing intelligent tutoring systems;
- behavioral experiments to identify good cognitive models of human learning processes.
- Applications of machine teaching to adversarial attacks, including cyber-security problems, generative adversarial networks, attacks on machine learning algorithms, etc.
- Novel applications for machine teaching such as program synthesis, human-robot interactions, social robotics, etc.
ScheduleAn updated schedule is available below (please refresh the page to see the latest version!).
- 0845 - 0900: Setting up the posters
- 0900 - 1000: Tutorial by Adish Singla, Jerry Zhu: "Introduction to Machine Teaching"
- 1000 - 1030: Talk by Emma Brunskill, Stanford - "Training the Trainers"
- 1030 - 1100: Coffee + Posters
- 1100 - 1130: Talk by Shay Moran, Technion - "Active Classification with Comparison Queries"
- 1130 - 1200: Discussion session (Topic: Open questions and new research directions)
Session 3 (1400 - 1630)
- 1400 - 1430: Talk by Le Song, Georgia Tech - "Iterative Machine Teaching"
- 1430 - 1500: Spotlight presentations for posters
- 1500 - 1530: Coffee + Posters
- 1530 - 1630: Poster session
- 1630 - 1700: Talk by Patrice Simard, Microsoft Research - "Machine Teaching: A New Paradigm for Building Machine Learning Systems"
- 1700 - 1730: Talk by Burr Settles, Duolingo - "Improving Language Learning and Assessment with Data"
- 1730 - 1800: Discussion session (Topic: Novel applications and industry insights)
- Generative Adversarial Active Learning
Jia-Jie Zhu; Jose Bento
- Machine Education - The Way Forward for Achieving Trust-Enabled Machine Agents
George Leu; Erandi Lakshika; Jiangjun Tang; Kathryn Merrick; Michael Barlow
- Optimizing Human Learning
Behzad Tabibian; Utkarsh Upadhyay; Abir De; Ali Zarezadeh; Bernhard Scholkopf; Manuel Gomez Rodriguez
- Model Distillation with Knowledge Transfer from Face Classification to Alignment and Verification
Chong Wang; Xipeng Lan
- Interpretable Machine Teaching via Feature Feedback
Shihan Su; Yuxin Chen; Oisin Mac Aodha; Pietro Perona; Yisong Yue
- Pedagogical Learning
Long Ouyang; Michael Frank
- Interpretable and Pedagogical Examples
Smitha L Milli; Pieter Abbeel; Igor Mordatch
- Faster Reinforcement Learning Using Active Simulators
Vikas Jain; Theja Tulabandula
- Gradual Tuning: A Better Way of Fine Tuning the Parameters of a Deep Neural Network
Guglielmo Montone; J.Kevin ORegan; Alexander V. Terekhov
- Accelerating Human Learning with Deep Reinforcement Learning
Siddharth Reddy; Anca Dragan; Sergey Levine
- Program2Tutor: Combining Automatic Curriculum Generation with Multi-Armed Bandits for Intelligent Tutoring Systems
Tong Mu; Karan Goel; Emma Brunskill
- Generative Knowledge Distillation for General Purpose Function Compression
Matthew Riemer; Michele Franceschini; Djallel Bouneffouf; Tim Klinger
- Explainable Artificial Intelligence via Bayesian Teaching
Scott Cheng-Hsin Yang; Patrick Shafto
- Predicting Recall Probability to Adaptively Prioritize Study
Shane Mooney; Karen Sun; Eric Bomgardner
- Machine Teaching: A New Paradigm for Building Machine Learning Systems
Patrice Y. Simard; Saleema Amershi; David M. Chickering; Alicia Edelman Pelton; Soroush Ghorashi; Christopher Meek; Gonzalo Ramos; Jina Suh; Johan Verwey; Mo Wang; John Wernsing