June 2008 @ Kobe, Japan

Kai-min Kevin Chang Research Associate (Special Faculty), Language Technologies Institute, School of Computer Science, Carnegie Mellon University. Affiliated with Center for Cognitive Brain Imaging, Center for the Neural Basis of Cognition Profile: CV, Resume Research Statement Email: kaimin dot chang at gmail dot com Phone: +1-412-268-1810 Fax: +1-412-268-6298 Office: Gates Hillman Center 5721 / Baker Hall 327M Address: Language Technologies Institute Carnegie Mellon University 5000 Forbes Avenue Pittsburgh, PA 15213 USA

Research Interests

My research interests include using machine learning and brain imaging technologies to investigate and model various human cognitive processes. My work includes theoretical contributions that study how is language encoded/decoded in the brain using functional Magnetic Resonance Imaging (fMRI), and practical applications of how do children learn to read using consumer-grade EEG headbands.

Recent advances in functional Magnetic Resonance Imaging (fMRI) provide a significant new approach to studying semantic representations in humans by making it possible to directly observe brain activity while people comprehend words, phrases, and even sentences. fMRI measures the hemodynamic response (changes in blood flow and blood oxygenation) related to neural activity in the human brain. Images can be acquired at good spatial resolution and reasonable temporal resolution - the activity level of 15,000 to 20,000 brain volume elements (voxels) of about 50 mm3 each can be measured every second. In previous research we produced and published the field's first predictive theory of neural representations of noun meanings (Mitchell et al., 2008; Chang et al., 2009), and demonstrated its success in predicting neural representations of arbitrary concrete nouns, across different human subjects. Building on this previous research, we are now investigating the influence of context (e.g., "doctor treated the patient" vs. "doctor drove the car") on the neural representation of a concept (e.g., "doctor"). Our goal is a general compositional modeling approach that will be applicable to predicting the neural representations of simple sentences, by determining, for example, how the neurosemantic properties of verbs and nouns combine.

The ultimate automated tutor could peer directly into students' minds to identify their mental states (knowledge, thoughts, feelings, and so forth) and decide accordingly what and how to teach at each moment. We proposed to investigate a novel source of input from as close to the brain as non-surgically practicable: EEG. The major goal of this project was to apply EEG technology in meaningful learning tasks (reading), using a uniquetestbed (Project LISTEN's Reading Tutor) to pursue useful targets (e.g., user intention, comprehension, if student is having difficulty, etc.). As part of this project, we collected ~3 years of tutor usage data collected in vivo at a primary school. The tutor is Project LISTEN's Reading Tutor and EEG was recorded with NeuroSky BrainBands. The Reading Tutor helps students learn how to read by listening (using Automated Speech Recognition) to them read story aloud. We annotated the time-course of a reading session with the sentence that the student was reading. The dataset consists of roughly 169 hours of EEG recording and 200,000 sentences. To assist researchers who are new to this topic, we also implemented a machine learning toolkit to help process the EEG data. We made both the (anonymized) dataset and toolkit publically available. Our notable results included using EEG to detect cheating (CSCW 2015), improve Knowledge Tracing (ITS 2014), detect comprehension (LAK 2014), detect engagement (AIED 2013), improve spoken dialog interface (ICMI 2012), improve automatic speech recognition (ACL 2012), improve intelligent tutoring system (IJAIED 2013).

Academic Activities

• Co-PIed on NSF 10-620 Cyberlearning: Tranforming Education grant (DIP: Exploiting Longitudinal EEG Input in a Reading Tutor), 2011-2014.
• Co-organized EDM 2015 Student Modeling Applications, Recent Developments & Toolkits (SMART) Tutorial, Madrid, Spain, 2015.
• Co-organized ITS 2014 Workshop on Utilizing EEG Input in Intelligent Tutoring Systems, Honolulu, HI, 2014.
• Co-organized NIPS 2014 Workshop on Machine Learning and Interpretation in Neuroimaging, Montreal, Canada, 2014.
• Co-organized NIPS 2013 Workshop on Machine Learning and Interpretation in Neuroimaging, Lake Tahoe, NV, 2013.
• Co-organized NIPS 2012 Workshop on Machine Learning and Interpretation in Neuroimaging, Lake Tahoe, NV, 2012.
• Co-organized NIPS 2011 Workshop on Machine Learning and Interpretation in Neuroimaging, Granada, Spain, 2011.
• Co-organized NAACL-HLT 2010 Workshop on Computational Neurolinguistics, Los Angeles, CA, 2010.
• Invited talk at Brain Informatics, Special Session on Computational Neurolinguistics, Toronto, Canada, 2010.

Software Tools

• Distribution of EEG-ML toolkit and Reading Tutor EEG data (Documentation, Download).
• Real-time EEG brain activity (MindSet).
• Distribution of the 60 Word-Pic fMRI data used in the Science paper (Science 2008 Data).
• 4CAPS Cognitive Neuroarchitecture (4CAPS).
• Bayes Net Toolbox for Student Modeling (BNT-SM).

Demo

• Synmetric (2-Minute Lessons Learned / Story Video, 1-Minute Technical Video).
• EEG-enabled MOOC (1).
• NeuroDialog (1, 2).
• Mind Reading (CBS 60 Minutes: Mind Reading).

Journal Publications

1. Jern, A., Chang, K.M., & Kemp, C. (2014). Belief polarization is not always irrational. Psychological Review, 121 (2), 206-224. (pdf)
2. Chang, K.M., Nelson, J., Pant, U., & Mostow, J. (2013). Toward Exploiting EEG Input in a Reading Tutor. International Journal of Artificial Intelligence in Education, 22 (1-2), 19-38. (pdf)
3. Chang, K.M., Mitchell, T.M., & Just M.A. (2010). Quantitative modeling of the neural representations of objects: How semantic feature norms can account for fMRI activation. NeuroImage: Special Issue on Multivariate Decoding and Brain Reading, 56, 716-727. (pdf, doi)
4. Kemp, C., Chang, K.M., & Lombardi, L. (2010). Category and feature identification. Acta Psychologica, 133, 216-233. (pdf)
5. Mitchell, T.M., Shinkareva, S.V., Carlson, A., Chang, K.M., Malave, V.L., Mason, R.A., & Just, M.A. (2008). Predicting human brain activity associated with noun meanings. Science, 320, 1191-1195. (pdf, supporting material)

Peer-Reviewed Conference Proceedings

1. Li, X., Chang, K.M., Yuan, Y., & Hauptmann, A. (2015). Massive Open Online Proctor: Protecting the Credibility of MOOCs certificates. Proceedings of the 18th ACM Conference on Computer Supported Cooperative Work & Social Computing, Vancouver, Canada, 1129-1137. (pdf)
2. Xu, Y., Chang, K.M., Yuan, Y., & Mostow, J. (2014). EEG Helps Knowledge Tracing! Proceedings of the 12th International Conference on Intelligent Tutoring Systems Workshop on Utilizing EEG Input in Intelligent Tutoring Systems, Honolulu, Hawaii. (pdf)
3. Yuan, Y., Chang, K.M., Xu, Y., & Mostow, J. (2014). A Public Toolkit and ITS Dataset for EEG. Proceedings of the 12th International Conference on Intelligent Tutoring Systems Workshop on Utilizing EEG Input in Intelligent Tutoring Systems, Honolulu, Hawaii. (pdf)
4. Xu, Y., Chang, K.M., Yuan, Y., & Mostow, J. (2014). Using EEG in Knowledge Tracing. The 7th International Conference on Educational Data Mining Workshop on Approaching Twenty Years of Knowledge Tracing: Lessons Learned, Open Challenges, and Promising Developments, London, UK.
5. Yuan, Y., Chang, K.M., Nelson, J., & Mostow, J. (2014). Toward Unobtrusive Measurement of Reading Comprehension Using Low-Cost EEG. Proceedings of the 4th International Conference on Learning Analytics and Knowledge, Indianapolis, Indiana. (pdf)
6. Wang, H., Li, Y. Hu, X., Yang, Y., Meng, Z., & Chang, K.M. (2013). Using EEG to Improve Massive Open Online Courses Feedback Interaction. Proceedings of the 16th International Conference on Artificial Intelligence in Education Workshop on Massive Open Online Courses, Memphis, Tennessee. (pdf)
7. Sridharan, S., Chen, Y.N., Chang, K.M., & Rudnicky, A. (2012). NeuroDialog: An EEG-Enabled Spoken Dialog Interface. Proceedings of the 14th ACM International Conference on Multi-modal Interaction, Santa Monica, CA, 65-66. (pdf)
8. Chen, Y.N., Chang, K.M., & Mostow, J. (2012) Towards using EEG to improve ASR accuracy. Proceedings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Montreal, Canada, 377-381. (pdf)
9. Chang, K.M., Murphy, B., & Just, M.A. (2012). A Latent Feature Analysis of the Neural Representation of Conceptual Knowledge. Proceedings of the 25th Advances in Neural Information Processing Systems Workshop on Machine Learning and Interpretation in Neuroimaging, Granada, Spain, Lecture Notes in Computer Science, 7263, 244-255. (pdf)
10. Mostow, J., Chang, K.M., & Nelson, J. (2011). Toward exploiting EEG input in a Reading Tutor. Proceedings of the 15th International Conference on Artificial Intelligence in Education, Auckland, New Zealand, 230-237. (pdf; Best Paper Nominee)
11. Chang, K.M., Cherkassky, V.L., Mitchell, T.M., & Just M.A. (2009). Quantitative modeling of the neural representation of adjective-noun phrases to account for fMRI activation. Proceedings of the 47th Annual Meeting of the ACL and the 4th IJCNLP of the AFNLP, Suntec, Singapore, 638-646. (pdf)
12. Jern, A., Chang, K.M., & Kemp, C. (2009). Bayesian belief polarization. Advances in Neural Information Processing Systems, Vancouver, Canada, 22. (pdf, supporting material)
13. Beck, J.E., Chang, K.M., Mostow, J., & Corbett, A. (2008). Does help help? Introducing the Bayesian evaluation and assessment methodology. Proceedings of the 9th International Conference on Intelligent Tutoring Systems, Montreal, Canada, 383-394. (pdf, Best Paper Award)
14. Beck, J.E. & Chang, K.M. (2007). Identifiability: A fundamental problem of student modeling. Proceedings of the 11th International Conference on User Modeling, Corfu, Greece, 137-146. (pdf)
15. Chang, K.M., Beck, J.E., Mostow, J., & Corbett, A. (2006). A Bayes net toolkit for student modeling in intelligent tutoring systems. Proceedings of the 8th International Conference on Intelligent Tutoring Systems, Jhongli, Taiwan, 104-113. (pdf)
16. Chang, K.M., Beck, J.E., Mostow, J., & Corbett, A. (2006). Does help help? A Bayes net approach to modeling tutor interventions. Proceedings of the 21st National Conference on Artificial Intelligence Workshop on Educational Data Mining, Boston, MA, 41-46. (pdf)
17. Chang, K.M., Beck, J.E., Mostow, J., & Corbett, A. (2005). Using speech recognition to evaluate two student models for a Reading Tutor. Proceedings of the 12th International Conference on Artificial Intelligence in Education Workshop on Student Modeling for Language Tutors, Amsterdam, Netherlands, 12-21. (pdf)
18. Beck, J.E., Chang, K.M., Mostow, J., & Corbett, A. (2005). Using a student model to improve a computer tutor's speech recognition. Proceedings of the 12th International Conference on Artificial Intelligence in Education Workshop on Student Modeling for Language Tutors, Amsterdam, Netherlands, 2-11. (pdf)

Poster Presentations

1. Register, C., Choi, S., Yuan, Y., Chang, K.M., & Mostow, J. (2014). Using Deep and Ensemble Machine Learning Algorithms to Decode the Brain. PSLC Summer Intern Poster Session, Carnegie Mellon University.
2. Han, S.J., Chang, K.M., Mostow, J. (2013). Exploiting Longitudinal EEG Input in a Reading Tutor: Detecting Natural Language Processing Features. Meeting of the Minds Undergraduate Research Symposium, Carnegie Mellon University.
3. Torres Olague, D., Yuan, Y., Chang, K.M., & Mostow, J. (2013). Can EEG detect when a student needs help? PSLC Summer Intern Poster Session, Carnegie Mellon University.
4. Kasman, E., Retterer-Moore, J., Xia, T., Nelson, J., Chang, K.M., & Mostow, J. (2012). How could brainwave information help an automated reading tutor? PSLC Summer Intern Poster Session, Carnegie Mellon University.
5. Pant, U., Chang, K.M., & Mostow, J. (2011). Testing takes time. Can EEG Detect Comprehension? PSLC Summer Intern Poster Session, Carnegie Mellon University.
6. Chang, K.M., Mitchell, T.M., & Just M.A. (2008). What functional brain imaging reveals about neuroarchitecture of object knowledge. The 30th Annual Meeting of the Cognitive Science Society, Washington, DC.
7. Chang, K.M., Malave V., Shinkareva, S., Mitchell, T.M., & Just M.A. (2007). What functional brain imaging reveals about neuroarchitecture of object knowledge. The 1st Okinawa Institute of Science and Technology Workshop on Cognitive Neurobiology, Okinawa, Japan.

Thesis

1. Chang, K.M. (2010). Quantitative modeling of the neural representation of nouns and phrases. Ph.D. Proposal. (pdf) (oral) Ph.D. Thesis. (pdf, oral)
2. Chang, K.M. (2003). Evolution of the dual route cascaded model of reading aloud. Undergraduate Thesis. (pdf)

Professional Activities

Biographical Sketch

Personal Interest

Some people write their diaries with words, some record them with pictures. I mark mine with food! Yes, I love to eat! My plan is taste all the savoury dishes in the world and mark them on my Savoury Google Maps! Still a long way to go, but I am getting there! :p I like to read the Slashdot, the tw.bbs.talk.joke newsgroup, and watch Comedy Central on TV. Three comic strips that I frequently visit are Piled Higher and Deeper, Dilbert, and River's 543. For leisure, I enjoy playing poker, chess and pool. I am also very into mobile devices. I frequent xda-developers and stay up to date on many smart phone devices. My current phone is AT & T Tilt2. Finally, I treasure freedom in speech, thoughts, codes, and am an advocate of Open Source software. PS, I was named a student of Watermelon according to this news article, originally published by University of Waterloo school officials on Apr 1, 2003. ;) Quite frankly, I joined Carnegie Mellon University later and that indeed made me a Watermelon. FYI, Kevyn Collins-Thompson is also a Watermelon.