CS372 Artificial Intelligence for Disease Diagnosis and Information Recommendations

Spring 2020 --- SARS-CoV-2 (COVID-19) Edition


3 units/18 sessions, including 15 lectures, and project proposal discussion and presentation.

Enrollment is limited to 30.


Edward Chang (echang@cs.stanford.edu)

Adjunct Professor, Computer Science, Stanford University

President, DeepQ Healthcare (areas: AI and healthcare)

Technical Advisor, SmartNews (areas: NLP and NLU)


TA: Saahil Jain (sj2675@cs.stanford.edu)





(3/18/20) Due to the COVID-19 pandemic, the 2020 edition of this course will focus on addressing COVID-19 prediction, containment, and mitigation

 (in addition to covering cancer, depression, and common diseases), using statistical models and AI methods.



      Artificial intelligence, specifically deep learning, stands out as one of the most transformative technologies of the past decade. AI can already outperform humans in several computer vision and natural language processing tasks.  However, we still face some of the same limitations and obstacles that led to the demise of the first AI boom phase five decades ago.  This research-oriented course will first review and reveal the limitations (e.g., iid assumption on training and testing data, voluminous training data requirement, and lacking interpretability) of some widely used AI algorithms, including convolutional neural networks (CNNs), transformers, reinforcement learning, and generative adversarial networks (GANs). To address these limitations, we will then explore topics including transfer learning for remedying data scarcity, knowledge-guided multimodal learning for improving data diversity, out of distribution generalization, attention mechanisms for enabling Interpretability, meta learning, and privacy-preserving training data management.  


The course will be taught through a combination of lecture and project sessions. Lectures on specialized AI applications (e.g., cancer/depression diagnosis and treatment, AI/VR for surgery, and health education) will feature guest speakers from academia and industry. Students will be assigned to work on an extensive project that is relevant to their fields of study (e.g., CS, Medicine, and Data Science). Projects may involve conducting literature surveys, formulating ideas, and implementing these ideas. Example project topics are but not limited to 1) knowledge guided GANs for improving training data diversity, 2) disease diagnosis via multimodal symptom checking, and 3) methods for COVID-19 path prediction and immunity analysis.


Level: CS senior/graduate, proficient in programming

Perquisite: Introductory course in AI, Statistics, or Machine Learning


Time and Location


·   Time: Tuesdays and Thursdays 9:00am to 10:20am

·   Location: On Zoom (McMurtry 360)




·      No exams.

·      Four short weekly assignments in the first half of the quarter.

·      Reading: A list of papers/articles/podcasts grouped by topics.  Each student selects two topics of interest to conduct in-depth surveys and presentation.

·      Term project: An extensive design and coding project.  Some example projects are:

1.      COVID-19 symptom checker and mitigator,

2.     A Virtual Assistant for news and/or COVID-19 information recommendations,

3.     COVID-19 outbreak tracking and prediction, and

4.     Methods for fusing knowledge and perception with deep learning.


Lecture Topics (Tentative, subject to minor changes)


Segment #1 (5 sessions)

What AI Can Do and what It Cannot Do

·      Intro to COVID-19

·      Deep Learning, Reinforcement Learning, and GANs

·      Multimodal BERT

·      AI Limitations, Obstacles, and Research Topics


Segment #2 (5 sessions)

Revealing Limitations with Case Studies in Healthcare

·      Sparse Data Reinforcement Learning for Symptom Checking

·      Transfer Learning and Multimodal Learning for Disease Diagnosis

·      COVID-19 Diagnosis, Containment, and Mitigation

·      Cancer Causes Treatment, 2 lectures (Ko, Stanford)

·      Depression/Anxiety Detection and Treatment (Duvvuri, Stanford MD)


Segment #3 (3 sessions)

Advanced Methods for Overcoming Limitations

·      Privacy Regulations and Measures for CCPA/GDPR Compliance

·      AI-assisted Programming and Debugging (Meijer, FB)

·      Successive Subspace Learning (SSL), Forward-Only Feature Learning for Robustness and Interpretability (Kuo, USC)


Segment #4 (2 sessions)

Tools: Virtual Assistant, HuggingFace, AutoML

·      Virtual Assistant (Giovanni, Stanford OVAL)

·      Dialogue and Query Refinement with NLP and RL (Stanford NLP group)


Term Projects (4 sessions)


Schedule (course materials will be updated when newer COVID-19 information is available, last updated 3/26/2020)



Course Materials




COVIS-19, what are we facing?  Contrarian opinions? What can AI help?


[1] The next outbreak? We’re not ready | Bill Gates [link], April 2015.

[2] What is a Virus, Vincent Racaniello,Columbia University [link].

[3] What is COVID-19 [link].

[4] Virus Classification, the Baltimore System [link].

[5] Coronavirus: Why You Must Act Now? [link].

[6] How to act? Cost and benefit of social distancing [link][link].

Assignment #1:

Virus ABC


4/14, 9am


COVID-19 containment and mitigation strategies, models and data sources


[1] The Science Behind the COVID-19, Patrick Soon-Shiong [link].

[2] Exponential growth and Epidemics [link].

[3] Estimating actual COVID-19 cases [link].

[4] COVID-19: Predicting the Path and Analyzing Immunity [link].

[5] IDM’s innovative software tools provide a quantitative and analytical means to model infectious disease [link].

[6] Impact of non-pharmaceutical interventions (NPIs) to reduce COVID19 mortality and healthcare demand, London Imperial College [link].

[7] How much ‘normal’ risk does COVID-19 represent? [link]

Assignment #2:

Analytical Model

and Data Sources


4/16, 9am


Artificial Intelligence in Medicine,

Past and Present


[1] Schwartz, W. B., R. S. Patil, and P. Szolovits. "AI in Medicine: where do we stand." New England Journal of Medicine 316 (1987): 685-688.

[2] Edward Y. Chang. "Artificial Intelligence in XPRIZE DeepQ Tricorder. " ACM MM Workshop for Personal Health and Health Care (2017).




AI for Symptom Checking


[1] Edward Y. Chang. "Context-Aware Symptom Checking for Disease Diagnosis Using Hierarchical Reinforcement Learning." AAAI (2018).

[2] Edward Y. Chang. "REFUEL: Exploring Sparse Features in Deep Reinforcement Learning for Fast Disease Diagnosis." NIPS (2018).

Assignment #3:

Term Project



4/21, In class


Term Project Discussion

Suggested Topics:

[1] COVID-19 & GDP joint prediction models for policy making

[2] Multimodal AI models for knowledge guided GANs

Organized by TA



Multimodal NLP: Word2vec, Transformers, BERT, GPT, Reformers, and beyond


[1] Attention Model

[2] BERT and GPT

[3] Reformers

[4] Multimodal BERT




Project Proposal Presentation

Each presentation should include 1) key objective, 2) related work,

3) proposed methods, and 4) data sources.


Organized by TA



Principles of Diagnosis


[1] Szolovits, P., and S. G. Pauker. "Categorical and Probabilistic Reasoning in Medical Diagnosis.” Artificial Intelligence 11(1-2), 1978: 115-144.

[2] Wu, T. D. "Efficient Diagnosis of Multiple Disorders Based on a Symptom Clustering Approach." Proceedings of AAAI, 1990, pp. 357-364.

[3] Patil, R. S., P. Szolovits, and W. B. Schwartz. "Causal Understanding of Patient Illness in Medical Diagnosis." In Proceedings of the Seventh International Joint Conference on Artificial Intelligence.




Cancer Causes and Diagnosis



Melissa Ko,

PhD, Cancer Biology,




Depression/Anxiety Diagnosis and Treatment



Dr. Vikas Duvvuri,

MD/PhD Stanford




Cancer Treatment and How AI May Help



Melissa Ko,

PhD, Cancer Biology,




Data Privacy and Regulations Compliance           



[2] Soteria: A Provably Compliant User Right Manager Using a Novel Two-Layer Blockchain Technology, Stanford University, DeepQ, NTU, 2020.




AI Assisted Coding and Debugging



Eric Meijer,



Interpretable and robust AI: Successive Subspace Learning (SSL), Forward-Only Feature Learning


[1] C.-C. Jay Kuo, “Understanding convolutional neural networks with a mathematical model,” the Journal of Visual Communications and Image Representation, Vol. 41, pp. 406-413, 2016.

[2] C.-C. Jay Kuo, Min Zhang, Siyang Li, Jiali Duan and Yueru Chen, “Interpretable Convolutional Neural Networks via Feedforward Design,” the Journal of Visual Communications and Image Representation, Vol. 60, pp. 346-359, 2019.

[3] M. Zhang, H. You, P. Kadam, S. Liu, C.-C. J. Kuo, “PointHop: an explainable machine learning method for point cloud classification,” arXiv:1907.12766 (2019), IEEE Trans. on Multimedia.

[4] Yueru Chen and C.-C. Jay Kuo, “PixelHop: a successive subspace learning (SSL) method for object classification,” arXiv:1909.08190 (2019), Visual Communications and Image Representation.

Prof. C.-C Jay Guo,




Term Project Presentation

Slots [1] [2] [3]

Organized by TA



Term Project Presentation

Slots [4] [5] [6]

Organized by TA



AI/VR Technologies for Brain Surgery & Medical Education


Panel: VR/AR for Surgery and Medical Education, April 2018 [link]

Stanford SNI Talk: Advancing Healthcare w/ AI and VR, Edward Y. Chang [link] (Stanford ID required)





TBD: Topics depend on the pandemic status in May