The AAAI organizing committee has decided that all sessions will be held as an In-Person event. The SafeAI workshop schedule is provided  below.

This year SafeAI has the honor and pleasure to announce our distinctive speakers Prof. Stuart Russell (University of California, Berkeley) and Prof. Song-Chun Zhu (PKU, THU, and BIGAI) who kindly accepted to conduct a special panel discussion, moderated by Mark Nitzberg (U.C. Berkeley), as opening and main event of the workshop. Special thanks to them!

Day 1: Scheduled on February 13, 2023 (Monday) from 09:00 to 17:10

Location: Room 202A, Walter E. Washington, Convention Center

Time (EST)

Description – 1st. Day Sessions

09:00-09:10 Welcome and Introduction – Workshop Chairs

Distinctive Panel Discussion: On Artificial General Intelligence and AI Alignment

Stuart Russell (University of California, Berkeley)

Song-Chun Zhu (Peking University, Tsinghua University, Beijing Institute for General Artificial Intelligence)

Chair: Mark Nitzberg (U.C. Berkeley)

10:30-11:00 Coffee Break & Posters Exhibition
Session 1: AI/ML Learning, Explainability, Accuracy and Policy Alignment – Chair: Gabriel Pedroza (CEA)

Active reward learning from multiple teachersPeter Barnett, Rachel Freedman, Justin Svegliato, Stuart Russell.
REVEALE: Reward Verification and Learning Using ExplanationsSaaduddin Mahmud, Sandhya Saisubramanian,  Shlomo Zilberstein.
A Robust Drift Detection Algorithm with High Accuracy and Low False Positives RateMaxime Fuccellaro, Laurent Simon, Akka Zemmari.

– Debate Panel – Session Discussants – Papers Authors and Chair

Session 2 – Short Presentations: AI/ML for Safety Critical Applications: Engineering, Evaluation, Monitoring – Chair: Mauricio Castillo-Effen (Lockheed Martin)

On Evaluating Adversarial Robustness of Chest X-ray Classification: Pitfalls & Best Practices – Yann Hicke, Salah Ghamizi , Maxime Cordy, Mike Papadakis and Yves Le Traon.
Towards Multi-timescale Online Monitoring of AI Models: Principles and Preliminary Results – Fateh Kaakai, Paul-Marie Raffi.
Capabilities for Better ML Engineering – Chenyang Yang, Rachel Brower-Sinning, Grace A. Lewis, Christian Kästner, Tongshuang Wu.

– Debate Panel – Session Discussants: Papers Authors and Chair, Sean McGregor (Responsible AI Collaborative).

12:20-14:00 Lunch Break
14:00-14:25 Invited Talk 1: Martin Rothfelder (Siemens), Digitalization and automation for driverless regional trains – The safe.trAIn research project
Session 3: AI/ML for Safety Critical Applications: Assurance Cases and Datasets – Chair: Gabriel Pedroza (CEA)

Transfer Assurance for Machine Learning in Autonomous Systems – Chiara Picardi, Richard Hawkins, Colin Paterson, Ibrahim Habli.
Domain-centric ADAS datasetsVáclav Diviš, Tobias Schuster, Marek Hrúz.
Towards Developing Safety Assurance Cases for Learning-Enabled Medical Cyber-Physical Systems – Maryam Bagheri, Josephine Lamp, Xugui Zhou, Lu Feng, Homa Alemzadeh.

– Debate Panel – Session Discussants – Papers Authors and Chair

Session 4 – Short Presentations: ML/DL Robustness: GAM and Attack Detection – Chair: Cynthia Chen (ETH Zurich)

Evaluation of GAN Architectures for Adversarial Robustness of Convolution Classifier – Qusay Mahmoud, Weimin Zhao, Sanaa Alwidian.
Backdoor Attack Detection in Computer Vision by applying Matrix Factorization on the the Weights of Deep Networks – Khondoker Hossain, Tim Oates.

– Debate Panel – Session Discussants: Papers Authors and Chair

15:35-16:05  Coffee Break and Posters Exhibition
Session 5: AI Safety Assessment: Failure-Cause Analysis, Assurance, Verification – Chair: Mauricio Castillo-Effen (Lockheed Martin)

A taxonomic system for failure cause analysis of open source AI incidents – Sean McGregor, Nikiforos Pittaras.
Towards Safety Assurance of Uncertainty-Aware Reinforcement Learning Agents – Karsten Roscher, Felippe Schmoeller da Roza, Simon Hadwiger, Ingo Thorn.
Formal Verification of Tree Ensembles against Real-World Composite Geometric Perturbations – Valency Colaco, Simin Nadjm-Tehrani.

– Debate Panel – Session Discussants – Papers Authors and Chair

17:00-17:10 Wrap Up 1st. Day – Workshop Chairs



Day 2: Scheduled on February 14 (Tuesday), 2023 from 09:30 to 17:15

Location: Room 202A, Walter E. Washington, Convention Center

Time (EST)

Description – 2nd. Day Sessions


Panel Discussion: Towards trustworthiness of AI-enabled systems: the Programme

Fateh Kaakai (Thales, IRT SystemX)

Souhaiel Khalfaoui (Valeo)

Augustin Lemesle (CEA)

Chair: Gabriel Pedroza (CEA)

10:30-11:00 Coffee Break & Posters Exhibition
Session 6: AI Robustness: Adversarial and Attacks Learning – Chair: Cynthia Chen (ETH Zurich)

Critically Assessing the State of the Art in CPU-based Local Robustness Verification – Matthias König, Annelot W. Bosman, Holger H. Hoos, Jan N. van Rijn.
Towards Understanding How Self-training Tolerates Data Backdoor Poisoning – Soumyadeep Pal, Ren Wang, Yuguang Yao, Sijia Liu.
Less is More: Data Pruning for Faster Adversarial Training – Yize Li, Pu Zhao, Xue Lin, Bhavya Kailkhura, Ryan Goldhahn
Personalized Models Resistant to Malicious Attacks for Human-centered Trusted AI – Kamil Kanclerz, Teddy Ferdinan, Jan Kocoń.

– Debate Panel – Session Discussants – Papers Authors and Chair, Khondoker Hossain (UMBC).

12:10-14:00 Lunch Break
14:00-14:25 Invited Talk 2: Vincent Conitzer (Carnegie Mellon University, University of Oxford), Foundations of Cooperative AI
Session 7: AI Robustness: Deep Reinforcement Learning – Chair: Mauricio Castillo-Effen (Lockheed Martin)

Robustness with Black-Box Adversarial Attack using Reinforcement Learning – Soumyendu Sarkar, Ashwin Ramesh Babu, Sajad Mousavi, Vineet Gundecha, Sahand Ghorbanpour, Alexander Shmakov, Ricardo Luna Gutierrez, Antonio Guillen, Avisek Naug.
White-Box Adversarial Policies in Deep Reinforcement Learning – Stephen Montes Casper, Dylan Hadfield-Menell, Gabriel Kreiman.
Bab: A novel algorithm for training clean model based on poisoned data – Chen Chen, Haibo Hong, Tao Xiang, Mande Xie, Jun Shao.
Safe Reinforcement Learning through Phasic Safety-Oriented Policy Optimization – Sumanta Dey, Pallab Dasgupta, Soumyajit Dey

– Debate Panel – Session Discussants – Papers Authors and Chair, Soumyendu Sarkar (Hewlett Packard), Khondoker Hossain (UMBC).

15:35-16:05  Coffee Break and Posters Exhibition
Session 8 – Short Presentations: OoD Detection and Uncertainty for ML/DL Safety – Chair: Mauricio Castillo-Effen (Lockheed Martin)

Out-of-Distribution Detection Using Deep Neural Network Latent Space Uncertainty – Fabio Arnez, Ansgar Radermacher, François Terrier.
Efficient and Effective Uncertainty Quantification in Gradient Boosting via Cyclical Gradient MCMC – Tian Tan, Carlos Huertas, Qi Zhao.
Safety Assurance with Ensemble-based Uncertainty Estimation and overlapping alternative Predictions in Reinforcement Learning – Dirk Eilers, Simon Burton, Felippe Schmoeller Roza, Karsten Roscher.

– Debate Panel – Session Discussants: Papers Authors and Chair, Axel Brando (Barcelona Supercomputing Center).

Session 9 – Short Presentations: Methods and Techniques for AI/ML Safety Assessment – Chair: Gabriel Pedroza (CEA)

Towards a holistic approach for AI trustworthiness assessment based upon aids for multi-criteria aggregation – Souhaiel Khalfaoui, Juliette Mattioli, Henri Sohier, Agnès Delaborde, Gabriel Pedroza, Kahina Amokrane-Ferka, Afef Awadid, Zakaria Chihani.
A Framework Quantifying Trustworthiness of Supervised Machine and Deep Learning Models – Pedro Miguel Sanchez, Alberto Huertas Celdran, Jan Kreischer, Melike Demirci, Joel Leupp, Muriel Figueredo Franco, Gérôme Bovet, Gregorio Martinez Perez, Burkhard Stiller.
Standardizing the Probabilistic Sources of Uncertainty for the sake of Safety Deep Learning – Axel Brando, Isabel Serra, Enrico Mezzetti, Francisco J. Cazorla, Jaume Abella.

– Debate Panel – Session Discussants: Papers Authors and Chair

16:55-17:15 Wrap Up & BPA – Workshop Chairs