Breaking Multi-Task Curse: Reward-Weighted Evolution for Black-Box Many-Task Optimization

Yanchi Li, Jiao Liu, Wenyin Gong, Qiong Gu, Yue Zhao, Yew-Soon Ong
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:68403-68429, 2026.

Abstract

Evolutionary multi-tasking accelerates black-box optimization via knowledge transfer but falters in scenarios involving many low-similarity tasks. We identify this scalability barrier as the Multi-Task Curse, driven by evaluation budget dispersion and negative transfer. To overcome this, we propose MES-RET (Many-task Evolution Strategy with Reward-weighted Evaluation and Transfer), which combats budget dispersion via a reward-weighted evaluation scheme that guarantees superior expected improvement, while simultaneously mitigating negative transfer through a robust reward-weighted aggregation of mean and covariance statistics, ensuring a safe fallback to independent evolution. Furthermore, to handle neural dimensional mismatches in many-task policy search, we introduce a semantic parameter alignment strategy that bridges heterogeneous state-action spaces. Extensive experiments on synthetic benchmarks, real-world engineering problems, and reinforcement learning tasks demonstrate that MES-RET consistently outperforms state-of-the-art methods, notably enabling skill transfer across morphologically distinct policies.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-li26dc, title = {Breaking Multi-Task Curse: Reward-Weighted Evolution for Black-Box Many-Task Optimization}, author = {Li, Yanchi and Liu, Jiao and Gong, Wenyin and Gu, Qiong and Zhao, Yue and Ong, Yew-Soon}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {68403--68429}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/li26dc/li26dc.pdf}, url = {https://proceedings.mlr.press/v306/li26dc.html}, abstract = {Evolutionary multi-tasking accelerates black-box optimization via knowledge transfer but falters in scenarios involving many low-similarity tasks. We identify this scalability barrier as the Multi-Task Curse, driven by evaluation budget dispersion and negative transfer. To overcome this, we propose MES-RET (Many-task Evolution Strategy with Reward-weighted Evaluation and Transfer), which combats budget dispersion via a reward-weighted evaluation scheme that guarantees superior expected improvement, while simultaneously mitigating negative transfer through a robust reward-weighted aggregation of mean and covariance statistics, ensuring a safe fallback to independent evolution. Furthermore, to handle neural dimensional mismatches in many-task policy search, we introduce a semantic parameter alignment strategy that bridges heterogeneous state-action spaces. Extensive experiments on synthetic benchmarks, real-world engineering problems, and reinforcement learning tasks demonstrate that MES-RET consistently outperforms state-of-the-art methods, notably enabling skill transfer across morphologically distinct policies.} }
Endnote
%0 Conference Paper %T Breaking Multi-Task Curse: Reward-Weighted Evolution for Black-Box Many-Task Optimization %A Yanchi Li %A Jiao Liu %A Wenyin Gong %A Qiong Gu %A Yue Zhao %A Yew-Soon Ong %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-li26dc %I PMLR %P 68403--68429 %U https://proceedings.mlr.press/v306/li26dc.html %V 306 %X Evolutionary multi-tasking accelerates black-box optimization via knowledge transfer but falters in scenarios involving many low-similarity tasks. We identify this scalability barrier as the Multi-Task Curse, driven by evaluation budget dispersion and negative transfer. To overcome this, we propose MES-RET (Many-task Evolution Strategy with Reward-weighted Evaluation and Transfer), which combats budget dispersion via a reward-weighted evaluation scheme that guarantees superior expected improvement, while simultaneously mitigating negative transfer through a robust reward-weighted aggregation of mean and covariance statistics, ensuring a safe fallback to independent evolution. Furthermore, to handle neural dimensional mismatches in many-task policy search, we introduce a semantic parameter alignment strategy that bridges heterogeneous state-action spaces. Extensive experiments on synthetic benchmarks, real-world engineering problems, and reinforcement learning tasks demonstrate that MES-RET consistently outperforms state-of-the-art methods, notably enabling skill transfer across morphologically distinct policies.
APA
Li, Y., Liu, J., Gong, W., Gu, Q., Zhao, Y. & Ong, Y.. (2026). Breaking Multi-Task Curse: Reward-Weighted Evolution for Black-Box Many-Task Optimization. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:68403-68429 Available from https://proceedings.mlr.press/v306/li26dc.html.

Related Material