AI & Machine Learning
Lucid Dreaming for Experience Replay
Reinforcement-learning agents typically learn from a static buffer of past transitions. Lucid Dreaming for Experience Replay (LiDER) “re-dreams” those memories under the agent’s current policy, keeping only refreshed trajectories that improve on the originals. The result is higher sample-efficiency and better Atari scores—all without altering the underlying off-policy algorithm.
Accountability & Legally-Aligned Fairness Metrics
When datasets are small or highly imbalanced, popular evaluation metrics can become mathematically unstable, masking bias. One study (AISTATS 2025) pinpoints this sample-size-induced bias and supplies reliability corrections. A companion effort (CIKM 2024) introduces the Objective Fairness Index (OFI), translating civil-rights doctrine into a robust numerical test. Together, these works give practitioners a principled toolkit for auditing models under real-world legal and data constraints.
Papers
- R.A. Rossi, D.F. Gleich and A.H. Gebremedhin, Parallel Maximum Clique Algorithms with Applications to Network Analysis, SIAM Journal on Scientific Computing, Vol 37, Issue 5, pages C589-C618, 2015.
Abstract Paper in PDF - B. Pattabiraman, M.M.A Patwary, A.H. Gebremedhin, W.K. Liao, A. Choudhary, Fast Algorithms for the Maximum Clique Problem on Massive Graphs with Applications to Overlapping Community Detection, Internet Mathematics, Vol 11, No 4-5, pp 421-448, 2015.
Abstract Paper in PDF - R.A. Rossi, D.F. Gleich, A.H. Gebremedhin and M.M.A Patwary, Fast Maximum Clique Algorithms for Large Graphs, Proceedings of WWW2014.
Abstract Paper in PDF - B. Pattabiraman, M.M.A Patwary, A.H. Gebremedhin, W.K. Liao, A. Choudhary, Fast Algorithms for the Maximum Clique Problem on Massive Sparse Graphs, WAW 2013: 10th Workshop on Algorithms and Models for the Web Graph, Lecture Notes in Computer Science 8305, pp 156-169, 2013.
Abstract Paper in PDF
- C. Soss, A. Rajam, J. Layne, E. Serra, M. Halappanavar, A. Gebremedhin.
ScaWL: Scaling k-WL (Weisfeiler-Leman) Algorithms in Memory and Performance on Shared and Distributed-Memory Systems, ACM TACO (accepted 2024).
Abstract | Paper in PDF
- X. Liu, M. Halappanavar, K. Barker, A. Lumsdaine, A. H. Gebremedhin.
Direction-Optimizing Label Propagation Framework for Structure Detection in Graphs: Design, Implementation, and Experimental Analysis, ACM Journal of Experimental Algorithmics 27 (1.12), 1–31 (2022).
Abstract | Paper in PDF - X. Liu, M. Halappanavar, K. Baker, A. Lumsdaine, A. H. Gebremedhin.
Direction-Optimizing Label Propagation and Its Application to Community Detection, Computing Frontiers 2020.
Abstract | Paper in PDF - X. Liu, J. Firos, M. Zalewski, M. Halappanavar, K. Baker, A. Lumsdaine, A. H. Gebremedhin.
Distributed Direction-Optimizing Label Propagation for Community Detection, IEEE HPEC 2019 (Graph Challenge Innovation Award).
Abstract | Paper in PDF - K. Sasani, MH. Namaki, Y. Wu, A.H. Gebremedhin, Multi-metric Graph Query Performance Prediction, 23rd International Conference on Database Systems for Advanced Applications (DASFAA 2018).
Abstract Paper in PDF - K. Sasani, MH. Namaki, A.H. Gebremedhin, Network Similarity Prediction in Time-evolving Graphs: A Machine Learning Approach, 32nd IEEE International Parallel and Distributed Processing Workshop on the Intersection of Graph Algorithms and Machine Learning (GraML 2018).
Abstract Paper in PDF - M.H. Namaki, K. Sasani, Y. Wu and A.H. Gebremedhin, Performance Prediction for Graph Queries, ACM SIGMOD International Conference on Management of Data Workshop on Network Data Analytics (NDA 2017).
Abstract Paper in PDF
- J. Stachofsky, A. H. Gebremedhin and R. Crossler.
Cast to Vote: A Socio-technical Network Analysis of an Election Smartphone Application, Digital Government: Research and Practice 3 (1), Article 3, 117 (2022).
Abstract | Paper in PDF
- X. Liu, J. Firoz, S. Aksoy, I. Amburg, A. Lumsdaine, C. Joslyn, B. Praggastis, A. H. Gebremedhin.
High-Order Line Graphs of Non-Uniform Hypergraphs: Algorithms, Applications, and Experimental Analysis, IEEE IPDPS 2022.
Abstract | Paper in PDF
- X. Liu, J. Firoz, A. H. Gebremedhin, A. Lumsdaine.
NWHy: A Framework for Hypergraph Analytics—Representations, Data Structures, and Algorithms, IEEE IPDPS Workshops 2022.
Abstract | Paper in PDF
- X. Liu, J. Firoz, A. Lumsdaine, C. Joslyn, S. Aksoy, B. Praggastis, A. H. Gebremedhin.
Parallel Algorithms for Efficient Computation of High-Order Line Graphs of Hypergraphs, IEEE HiPC 2021.
Abstract | Paper in PDF
- S. Ghosh, M. Halappanavar, A. Kalyanaraman, A. Khan, A. H. Gebremedhin.
Exploring MPI Communication Models for Graph Applications Using Graph Matching as a Case Study, IEEE IPDPS 2019.
Abstract | Paper in PDF