schedule overview

How to access

To access the Zoom link, visit the IJCAI Gather.Town website (https://ijcai-21.org/ijcai-virtual-venue-access/), then 'walk' to Montreal Green, and then walk to room Green 2. Press X to see the link.

The poster sessions are in the Gather.Town space, outside of room "Green 2". At the box "Press X for directory", press x, then you can see the locations of posters of three sessions (Session W38_1, Session W38_2, Session W38_3).

Coffee breaks are in the Gather.Town space, where you can talk in small 'groups' like in a real workshop.

Detailed program

All talks are 7m presentation + 3m Q&A, followed by a poster.

19 August (14.00-20.00), UTC

14:00-15:10 Session 1

  • Learning Deeper Variable Ordering Heuristics for Constraint Optimisation, by Floris Doolaard and Neil Yorke-Smith (paper)

  • Toward Learning Mixed-Integer Linear Programs from Contextual Examples, by Mohit Kumar, Samuel Kolb, Luc De Raedt and Stefano Teso (paper)

  • Learning Interpretable Error Functions for Combinatorial Optimization Problem Modeling, by Florian Richoux and Jean-Francois Baffier (paper)

  • Active Learning Meets Optimized Item Selection, by Bernard Kleynhans, Xin Wang, Serdar Kadıoğlu (paper)

  • Automated personnel rostering with implicit constraints, by Pieter Smet, Robbe De Rijck, Tony Wauters (paper)

  • Injecting Domain Knowledge in Neural Networks: a Controlled Experiment on a Constrained Problem (Extended Abstract), by Mattia Silvestri, Michele Lombardi, Michela Milano (paper)

  • Ensuring the Quality of Optimization Solutions in Data Generated Optimization Models, by Segev Wasserkrug, Orit Davidovich, Evegeny Shindin, Dharmashankar Subramanian, Parikshit Ram, Pavankumar Murali, Dzung Phan, Nianjun Zhou and Lam M. Nguyen (paper)


15:10-16:10 Posters

16:10-16:20 break

16:20-17:20 Keynote: Learning, Optimization, and Generalization in the Predict-then-Optimize Setting,
Paul Grigas (University of California, Berkeley)

17:20-17:35 break

17:35-18:45 Session 2

  • Teaching the Old Dog New Tricks: Supervised Learning with Constraints, by Fabrizio Detassis, Michele Lombardi, Michela Milano (paper)

  • Lessons from the Clustering Analysis of a Search Space: A Centroid-based Approach to Initializing NAS, by Kalifou Rene Traore, Andres Camero, Xiao Xiang Zhu (paper)

  • Dual Evolutionary Ensemble Learning: An Automatic Ensemble Classifier Construction Method, by Hao Chen, Guo Xin Zhang, Xiao Ying Pan and Rong Jia (paper)

  • Chance constrained conic-segmentation support vector machine with uncertain data, by Shen Peng and Gianpiero Canessa (paper)

  • GradFreeBits: Gradient Free Bit Allocation for Dynamic Low Precision Neural Networks, by Benjamin Bodner, Gil Ben Shalom, Eran Treister (paper)

  • Hidden State Approximation in Recurrent Neural Networks Using Continuous Particle Filtering, by Dexun Li, Pradeep Varakantham (paper)

  • Deep Ensemble Networks for the identification of Parasitized Malaria from Blood Smear Cellular Images, by Dipam Paul, Alankrita Tewari (paper)


18:45-19:45 Posters

20 August (14.00-20.00), UTC

14:00-15:20 Session 3

  • Deep Reinforcement Learning with Hand-crafted Features to Solve P-median Problems, by En-Cheng Chang, Paula Carroll, Deepak Ajwani (paper)

  • A Learning and Optimization Framework for Collaborative Urban Delivery Problem with Alliances, by Jingfeng Yang and Hoong Chuin Lau (paper)

  • Learning from Obstructions: An Effective Deep Learning Approach for Minimum Vertex Cover, by Faisal N. Abu-Khzam, Mohamed M. Abd El-Wahab, Moussa Haidous, Noureldin Yosri (paper)

  • Deep Reinforcement Learning and Optimization Approach for Multi-echelon Supply Chain with Uncertain Demands, by Julio C´esar Alves and Geraldo Robson Mateus (paper)

  • Integrating Deep Learning and Optimization for Epidemic Control, by Federico Baldo, Andrea Borghesi, Michela Milano (paper)

  • Local Search and Machine Learning for Capacitated Vehicle Routing (work in progress paper), by Teun Druijf, Ad Feelders and Han Hoogeveen (paper)

  • Experiments with graph convolutional networks for solving the vertex p-center problem, by Elisabeth Gaar and Markus Sinnl (paper)

  • A general forecasting-based portfolio optimization model, by Thibault Lechien , Jens Goemaere and Patrick De Causmaecker (paper)


15:20-16:20 Poster

16:20-16:30 break

16:30-17:30 Keynote: The opportunity of AI – How AI can improve or even replace traditional Optimization
Patrick Hennen (ORTEC)

17:30-17:45 break

17:45-19:20 Session 4: talks from AI for TSP competition, session chair: Paulo da Costa

  • 17:45 -18:00: Introduction and winner announcement, by Laurens Bliek

    • 18:00-18:30: Track 1
      Team ZLI (Manuel López-Ibáñez, Ekhine Irurozki Arrieta, Martin Zaefferer)
      A Self-Adaptive Bayesian Optimizer based on Clustered Kriging and Feasibility Classification for the Black-box Time-Dependent Orienteering Problem with Stochastic Weights and Time Windows
      Team Convexers (Meinolf Sellmann, Tapan Shah, Kevin Tierney)
      Hyper-configurable reactive search with surrogates
      Team Margaridinhas (Warley Almeida, Federico Bobbio, Caroline Leboeuf, Justine Pepin, Carl Perreault-Lafleur)
      An iterative approach based on a mixed-integer surrogate model for the time-dependent orienteering traveling salesman problem with stochastic weights and time windows

    • 18:30 - 19:00: Track 2
      Team Rise up (Fynn Schmitt-Ulms, André Hottung, Kevin Tierney, Meinolf Sellmann)
      Solving Stochastic Routing Problems with POMO and Efficient Active Search
      Team Ratel (Ricardo Gama, Hugo L. Fernandes)
      A recursive attentive Pointer Network model for the OPSWTW: a reinforcement learning approach

  • 19:00 - 19:20: discussions

19:20 - 20:00 Social time