CS 6463 Advanced Topics: Trust in AI
Instructor:
Sumit Kumar Jha
Optional
Meeting: Friday 3:30 pm @ Zoom
Part I: Explainable Artificial Intelligence
Week 1-2 (Jan 17 to Jan 27): Deep Inside
Convolutional Networks: Visualizing Image Classification Models and Saliency
Maps https://arxiv.org/pdf/1312.6034.pdf
Week 3 (Jan 30 to Feb 3): Grad-CAM: Visual
Explanations from Deep Networks via Gradient-based Localization https://arxiv.org/pdf/1610.02391.pdf
Week 4 (Feb 6 to Feb 10): Axiomatic
Attribution for Deep Networks (Integrated Gradients) http://proceedings.mlr.press/v70/sundararajan17a/sundararajan17a.pdf
Week 5 (Feb 13 to Feb 17): SmoothGrad:
removing noise by adding noise https://arxiv.org/pdf/1706.03825.pdf
Week 6 (Feb 20 to Feb 24): Fast Axiomatic
Attribution for Neural Networks https://arxiv.org/pdf/2111.07668.pdf
Week 7: (Feb 27 to March 3)
Self-Attention Attribution: Interpreting Information Interactions Inside
Transformer https://arxiv.org/abs/2004.11207
Lecture
and Related Materials
Part II: Confidence Metrics for
Neural Networks
Week 8 (March 6 to March 10) : Probabilistic Outputs for Support Vector Machines and
Comparisons to Regularized Likelihood Methods https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.41.1639
Lecture
and Related Materials
Week 9: Spring Break
Week 10 (March 20 to March 24): On
calibration of modern neural networks https://proceedings.mlr.press/v70/guo17a/guo17a.pdf
Lecture
and Related Materials
Week 11 (March 27 to 31): Attribution-Based
Confidence Metric For Deep Neural Networks
https://proceedings.neurips.cc/paper/2019/file/bc1ad6e8f86c42a371aff945535baebb-Paper.pdf
Lecture
and Related Materials
Part III: NeuroSymbolic
AI
Week: D-CODE: Discovering Closed-form ODEs from Observed Trajectories https://openreview.net/forum?id=wENMvIsxNN�
Lecture and Related Materials
Part IV: Large Visual and Language Models
Week 12
(April 3 to 7): TBD
Week 13
(April 10 to 14): TBD
Week 14
(April 17 to 21): TBD
Week 15
(April 24 to 28): TBD
Week 16 (May
1 to 2): In-class Project (optional) Presentations (videos available on UTSA
Blackboard/Canvas).
May 2, 2023: Last Day of Classes