Eliciting good teaching from humans for machine learners

We propose using computational teaching algorithms to improve human teaching for machine learners. We investigate example sequences produced naturally by human teachers and find that humans often do not spontaneously generate optimal teaching sequences for arbitrary machine learners. To elicit better teaching, we propose giving humans teaching guidance, which are instructions on how to teach, derived from computational teaching algorithms or heuristics. We present experimental results demonstrating that teaching guidance substantially improves human teaching in three different problem domains. This provides promising evidence that human intelligence and flexibility can be leveraged to achieve better sample efficiency when input data to a learning system comes from a human teacher. We aim to improve Interactive Machine Learning by influencing the human teacher.We propose Teaching Guidance: instructions for teachers, to improve their input.Teaching Guidance is derived from optimal or heuristic teaching algorithms.We performed experiments to compare human teaching with and without teaching guidance.We found that Teaching Guidance substantially improves the data provided by teachers.

[1]  Stephanie Rosenthal,et al.  Towards maximizing the accuracy of human-labeled sensor data , 2010, IUI '10.

[2]  Weng-Keen Wong,et al.  Integrating rich user feedback into intelligent user interfaces , 2008, IUI '08.

[3]  Laura A. Dabbish,et al.  Labeling images with a computer game , 2004, AAAI Spring Symposium: Knowledge Collection from Volunteer Contributors.

[4]  Peter Stone,et al.  A social reinforcement learning agent , 2001, AGENTS '01.

[5]  Jason Weston,et al.  Curriculum learning , 2009, ICML '09.

[6]  Desney S. Tan,et al.  CueTIP: a mixed-initiative interface for correcting handwriting errors , 2006, UIST.

[7]  Paul A. Viola,et al.  Interactive Information Extraction with Constrained Conditional Random Fields , 2004, AAAI.

[8]  Rémi Gilleron,et al.  PAC Learning under Helpful Distributions , 1997, RAIRO Theor. Informatics Appl..

[9]  Thomas G. Dietterich,et al.  Interacting meaningfully with machine learning systems: Three experiments , 2009, Int. J. Hum. Comput. Stud..

[10]  Manuela M. Veloso,et al.  Layered Learning , 2000, ECML.

[11]  Christoph H. Lampert,et al.  Learning to detect unseen object classes by between-class attribute transfer , 2009, 2009 IEEE Conference on Computer Vision and Pattern Recognition.

[12]  Thomas Zeugmann,et al.  Recent Developments in Algorithmic Teaching , 2009, LATA.

[13]  Oscar Déniz-Suárez,et al.  Learning to Recognize Faces Incrementally , 2007, DAGM-Symposium.

[14]  Peter Stone,et al.  Cobot: A Social Reinforcement Learning Agent , 2001, NIPS.

[15]  A. Thomaz,et al.  Mixed-Initiative Active Learning , 2012 .

[16]  Noah D. Goodman,et al.  Teaching Games : Statistical Sampling Assumptions for Learning in Pedagogical Situations , 2008 .

[17]  Michael Kearns,et al.  On the complexity of teaching , 1991, COLT '91.

[18]  Andrea L. Thomaz,et al.  Socially guided machine learning , 2006 .

[19]  James N. MacGregor The Effects of Order on Learning Classifications by Example: Heuristics for Finding the Optimal Order , 1988, Artif. Intell..

[20]  Bilge Mutlu,et al.  How Do Humans Teach: On Curriculum Learning and Teaching Dimension , 2011, NIPS.

[21]  Andrea Lockerd Thomaz,et al.  Reinforcement Learning with Human Teachers: Evidence of Feedback and Guidance with Implications for Learning Performance , 2006, AAAI.

[22]  Bill Tomlinson,et al.  Who are the crowdworkers?: shifting demographics in mechanical turk , 2010, CHI Extended Abstracts.

[23]  H. David Mathias,et al.  A Model of Interactive Teaching , 1997, J. Comput. Syst. Sci..

[24]  Sanjoy Dasgupta,et al.  Coarse sample complexity bounds for active learning , 2005, NIPS.

[25]  N. Littlestone Learning Quickly When Irrelevant Attributes Abound: A New Linear-Threshold Algorithm , 1987, 28th Annual Symposium on Foundations of Computer Science (sfcs 1987).

[26]  Desney S. Tan,et al.  CueFlik: interactive concept learning in image search , 2008, CHI.

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

[28]  Herman Chernoff,et al.  The Use of Faces to Represent Points in k- Dimensional Space Graphically , 1973 .

[29]  J. Tenenbaum,et al.  A global geometric framework for nonlinear dimensionality reduction. , 2000, Science.

[30]  Andreas Krause,et al.  Near-Optimally Teaching the Crowd to Classify , 2014, ICML.

[31]  Dana Angluin,et al.  Queries and concept learning , 1988, Machine Learning.

[32]  Aniket Kittur,et al.  Crowdsourcing user studies with Mechanical Turk , 2008, CHI.

[33]  M. Mascolo Change processes in development: The concept of coactive scaffolding , 2005 .

[34]  J. Elman Learning and development in neural networks: the importance of starting small , 1993, Cognition.

[35]  Paul A. Viola,et al.  Corrective feedback and persistent learning for information extraction , 2006, Artif. Intell..

[36]  Jerry Alan Fails,et al.  A design tool for camera-based interaction , 2003, CHI '03.

[37]  Todd Kulesza,et al.  Tell me more?: the effects of mental model soundness on personalizing an intelligent agent , 2012, CHI.

[38]  Tom M. Mitchell,et al.  Text clustering with extended user feedback , 2006, SIGIR.

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

[40]  Matthew E. Taylor Assisting Transfer-Enabled Machine Learning Algorithms: Leveraging Human Knowledge for Curriculum Design , 2009, AAAI Spring Symposium: Agents that Learn from Human Teachers.

[41]  Peter Stone,et al.  Training a Tetris agent via interactive shaping: a demonstration of the TAMER framework , 2010, AAMAS.

[42]  Manuel Lopes,et al.  Algorithmic and Human Teaching of Sequential Decision Tasks , 2012, AAAI.

[43]  Garrison W. Cottrell,et al.  A stochastic optimal control perspective on affect- sensitive teaching , 2012 .

[44]  James L. McClelland,et al.  Success and failure in teaching the [r]-[l] contrast to Japanese adults: Tests of a Hebbian model of plasticity and stabilization in spoken language perception , 2002, Cognitive, affective & behavioral neuroscience.

[45]  Patrick Shafto,et al.  Reasoning in teaching and misleading situations , 2011, CogSci.

[46]  Yong Jae Lee,et al.  Learning the easy things first: Self-paced visual category discovery , 2011, CVPR 2011.