Papers

For a full list of my papers, please see my Google Scholar profile.

Chen L, Chen J, Liu C, Kirchenbauer J, Soselia D, Zhu C, Goldstein T, Huang H (2024). OPTune: Efficient Online Preference Tuning. arXiv preprint arXiv:2406.07657.

Chen L, Zhu C, Soselia D, Chen J, Zhou T, Goldstein T, Huang H (2024). Odin: Disentangled Reward Mitigates Hacking in RLHF. Proceedings of the 41st International Conference on Machine Learning (ICML 2024).

Ding P, Soselia D, Armstrong T, Su J, Huang F (2023). Reviving Shift Equivariance in Vision Transformers. The Second Workshop on Spurious Correlations, Invariance and Stability (SCIS), ICML 2023.

Soselia D, Saifullah K, Zhou T (2023). Learning UI-to-Code Reverse Generator Using Visual Critic Without Rendering. arXiv preprint arXiv:2305.14637.

Zhang L, Soselia D, Wang R, Gutierrez-Farewik EM (2023). Estimation of Joint Torque by EMG-Driven Neuromusculoskeletal Models and LSTM Networks. IEEE Transactions on Neural Systems and Rehabilitation Engineering.

Zhang L, Soselia D, Wang R, Gutierrez-Farewik EM (2022). Lower-Limb Joint Torque Prediction Using LSTM Neural Networks and Transfer Learning. IEEE Transactions on Neural Systems and Rehabilitation Engineering.

Soselia Davit (2022). Continual Learning and Biomedical Image Data: Attempting to sequentially learn medical imaging datasets using continual learning approaches.

Olsson F, Soselia D, Hijmans JM, Gutierrez-Farewik EM (2021). Predicting Rocker Shoe Parameters for Reducing Plantar Pressure Using Machine Learning Methods. Gait & Posture, Elsevier.

Paolini C, Soselia D, Baweja H, Sarkar M (2019). Optimal Location for Fall Detection Edge Inferencing. 2019 IEEE Global Communications Conference (GLOBECOM).

Soselia D, Shugliashvili L, Amashukeli S, Koberidze I, Chelidze G (2018). Freezing Networks: Weight Preservation Procedure for Continual Learning. NeurIPS Continual Learning Workshop.

Soselia D, Tsintsadze M, Shugliashvili L, Koberidze I, Amashukeli S (2018). On Georgian Handwritten Character Recognition. IFAC-PapersOnLine. 51.0(30.0): 161-165. Elsevier.