Ondrej Bohdal

41 papers A* 8A 2Journal 29Unranked 2
YearRankTypeTitle / Venue / Authors
2026 J jnl
Clustering-driven Memory Compression for On-device Large Language Models.
CoRR
Ondrej Bohdal, Pramit Saha, Umberto Michieli, Mete Ozay, Taha Ceritli
2026 J jnl
Data-driven Clustering and Merging of Adapters for On-device Large Language Models.
CoRR
Ondrej Bohdal, Taha Ceritli, Mete Ozay, Jijoong Moon, Kyeng-Hun Lee, Hyeonmok Ko, Umberto Michieli
2026 J jnl
Diffusion Alignment Beyond KL: Variance Minimisation as Effective Policy Optimiser.
CoRR
Zijing Ou, Jacob Si, Junyi Zhu, Ondrej Bohdal, Mete Ozay, Taha Ceritli, Yingzhen Li
2025 J jnl
CG-TTRL: Context-Guided Test-Time Reinforcement Learning for On-Device Large Language Models.
CoRR
Peyman Hosseini, Ondrej Bohdal, Taha Ceritli, Ignacio Castro, Matthew Purver, Mete Ozay, Umberto Michieli
2025 A* conf
Efficient Compositional Multi-tasking for On-device Large Language Models.
EMNLP
Ondrej Bohdal, Mete Ozay, Jijoong Moon, Kyeng-Hun Lee, Hyeonmok Ko, Umberto Michieli
2025 J jnl
Efficient Compositional Multi-tasking for On-device Large Language Models.
CoRR
Ondrej Bohdal, Mete Ozay, Jijoong Moon, Kyeng-Hun Lee, Hyeonmok Ko, Umberto Michieli
2025 A* conf
HydraOpt: Navigating the Efficiency-Performance Trade-off of Adapter Merging.
EMNLP
Taha Ceritli, Ondrej Bohdal, Mete Ozay, Jijoong Moon, Kyeng-Hun Lee, Hyeonmok Ko, Umberto Michieli
2025 J jnl
HydraOpt: Navigating the Efficiency-Performance Trade-off of Adapter Merging.
CoRR
Taha Ceritli, Ondrej Bohdal, Mete Ozay, Jijoong Moon, Kyeng-Hun Lee, Hyeonmok Ko, Umberto Michieli
2025 J jnl
K-Merge: Online Continual Merging of Adapters for On-device Large Language Models.
CoRR
Donald Shenaj, Ondrej Bohdal, Taha Ceritli, Mete Ozay, Pietro Zanuttigh, Umberto Michieli
2025 A conf
MemControl: Mitigating Memorization in Diffusion Models via Automated Parameter Selection.
WACV
Raman Dutt, Ondrej Bohdal, Pedro Sanchez, Sotirios A. Tsaftaris, Timothy M. Hospedales
2025 J jnl
MemLoRA: Distilling Expert Adapters for On-Device Memory Systems.
CoRR
Massimo Bini, Ondrej Bohdal, Umberto Michieli, Zeynep Akata, Mete Ozay, Taha Ceritli
2025 — conf
On-device System of Compositional Multi-tasking in Large Language Models.
EMNLP (Industry Track)
Ondrej Bohdal, Konstantinos Theodosiadis, Asterios Mpatziakas, Dimitrios Filippidis, Iro Spyrou, Christos Zonios, Anastasios Drosou, Dimosthenis Ioannidis, Kyeng-Hun Lee, Jijoong Moon, Hyeonmok Ko, Mete Ozay, Umberto Michieli
2025 J jnl
On-device System of Compositional Multi-tasking in Large Language Models.
CoRR
Ondrej Bohdal, Konstantinos Theodosiadis, Asterios Mpatziakas, Dimitris Filippidis, Iro Spyrou, Christos Zonios, Anastasios Drosou, Dimosthenis Ioannidis, Kyeng-Hun Lee, Jijoong Moon, Hyeonmok Ko, Mete Ozay, Umberto Michieli
2025 A* conf
VL-ICL Bench: The Devil in the Details of Multimodal In-Context Learning.
ICLR
Yongshuo Zong, Ondrej Bohdal, Timothy M. Hospedales
2024 J jnl
Capacity Control is an Effective Memorization Mitigation Mechanism in Text-Conditional Diffusion Models.
CoRR
Raman Dutt, Pedro Sanchez, Ondrej Bohdal, Sotirios A. Tsaftaris, Timothy M. Hospedales
2024 A* conf
FairTune: Optimizing Parameter Efficient Fine Tuning for Fairness in Medical Image Analysis.
ICLR
Raman Dutt, Ondrej Bohdal, Sotirios A. Tsaftaris, Timothy M. Hospedales
2024 A conf
Feed-Forward Latent Domain Adaptation.
WACV
Ondrej Bohdal, Da Li, Shell Xu Hu, Timothy M. Hospedales
2024 J jnl
LoRA.rar: Learning to Merge LoRAs via Hypernetworks for Subject-Style Conditioned Image Generation.
CoRR
Donald Shenaj, Ondrej Bohdal, Mete Ozay, Pietro Zanuttigh, Umberto Michieli
2024 J jnl
MemControl: Mitigating Memorization in Medical Diffusion Models via Automated Parameter Selection.
CoRR
Raman Dutt, Pedro Sanchez, Ondrej Bohdal, Sotirios A. Tsaftaris, Timothy M. Hospedales
2024 J jnl
Memorized Images in Diffusion Models share a Subspace that can be Located and Deleted.
CoRR
Ruchika Chavhan, Ondrej Bohdal, Yongshuo Zong, Da Li, Timothy M. Hospedales
2024 J jnl
Navigating Noise: A Study of How Noise Influences Generalisation and Calibration of Neural Networks.
Trans. Mach. Learn. Res.
Martin Ferianc, Ondrej Bohdal, Timothy M. Hospedales, Miguel R. D. Rodrigues
2024 A* conf
Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models.
ICML
Yongshuo Zong, Ondrej Bohdal, Tingyang Yu, Yongxin Yang, Timothy M. Hospedales
2024 J jnl
Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models.
CoRR
Yongshuo Zong, Ondrej Bohdal, Tingyang Yu, Yongxin Yang, Timothy M. Hospedales
2024 J jnl
VL-ICL Bench: The Devil in the Details of Benchmarking Multimodal In-Context Learning.
CoRR
Yongshuo Zong, Ondrej Bohdal, Timothy M. Hospedales
2023 J jnl
FairTune: Optimizing Parameter Efficient Fine Tuning for Fairness in Medical Image Analysis.
CoRR
Raman Dutt, Ondrej Bohdal, Sotirios A. Tsaftaris, Timothy M. Hospedales
2023 J jnl
Fairness in AI and Its Long-Term Implications on Society.
CoRR
Ondrej Bohdal, Timothy M. Hospedales, Philip H. S. Torr, Fazl Barez
2023 J jnl
Feed-Forward Source-Free Domain Adaptation via Class Prototypes.
CoRR
Ondrej Bohdal, Da Li, Timothy M. Hospedales
2023 J jnl
Impact of Noise on Calibration and Generalisation of Neural Networks.
CoRR
Martin Ferianc, Ondrej Bohdal, Timothy M. Hospedales, Miguel R. D. Rodrigues
2023 J jnl
Label Calibration for Semantic Segmentation Under Domain Shift.
CoRR
Ondrej Bohdal, Da Li, Timothy M. Hospedales
2023 A* conf
Meta Omnium: A Benchmark for General-Purpose Learning-to-Learn.
CVPR
Ondrej Bohdal, Yinbing Tian, Yongshuo Zong, Ruchika Chavhan, Da Li, Henry Gouk, Li Guo, Timothy M. Hospedales
2023 J jnl
Meta Omnium: A Benchmark for General-Purpose Learning-to-Learn.
CoRR
Ondrej Bohdal, Yinbing Tian, Yongshuo Zong, Ruchika Chavhan, Da Li, Henry Gouk, Li Guo, Timothy M. Hospedales
2023 J jnl
Meta-Calibration: Learning of Model Calibration Using Differentiable Expected Calibration Error.
Trans. Mach. Learn. Res.
Ondrej Bohdal, Yongxin Yang, Timothy M. Hospedales
2023 A* conf
PASHA: Efficient HPO and NAS with Progressive Resource Allocation.
ICLR
Ondrej Bohdal, Lukas Balles, Martin Wistuba, Beyza Ermis, Cédric Archambeau, Giovanni Zappella
2022 J jnl
Feed-Forward Source-Free Latent Domain Adaptation via Cross-Attention.
CoRR
Ondrej Bohdal, Da Li, Shell Xu Hu, Timothy M. Hospedales
2022 J jnl
PASHA: Efficient HPO with Progressive Resource Allocation.
CoRR
Ondrej Bohdal, Lukas Balles, Beyza Ermis, Cédric Archambeau, Giovanni Zappella
2021 — conf
A Channel Coding Benchmark for Meta-Learning.
NeurIPS Datasets and Benchmarks
Rui Li, Ondrej Bohdal, Rajesh K. Mishra, Hyeji Kim, Da Li, Nicholas D. Lane, Timothy M. Hospedales
2021 J jnl
A Channel Coding Benchmark for Meta-Learning.
CoRR
Rui Li, Ondrej Bohdal, Rajesh K. Mishra, Hyeji Kim, Da Li, Nicholas D. Lane, Timothy M. Hospedales
2021 A* conf
EvoGrad: Efficient Gradient-Based Meta-Learning and Hyperparameter Optimization.
NeurIPS
Ondrej Bohdal, Yongxin Yang, Timothy M. Hospedales
2021 J jnl
EvoGrad: Efficient Gradient-Based Meta-Learning and Hyperparameter Optimization.
CoRR
Ondrej Bohdal, Yongxin Yang, Timothy M. Hospedales
2021 J jnl
Meta-Calibration: Meta-Learning of Model Calibration Using Differentiable Expected Calibration Error.
CoRR
Ondrej Bohdal, Yongxin Yang, Timothy M. Hospedales
2020 J jnl
Flexible Dataset Distillation: Learn Labels Instead of Images.
CoRR
Ondrej Bohdal, Yongxin Yang, Timothy M. Hospedales