# Theoretical Investigations and Practical Enhancements on Tail Task Risk Minimization in Meta Learning

Source: [Theoretical Investigations and Practical Enhancements on Tail Task Risk Minimization in Meta Learning](<https://www.sota2.com/research/paper/68fa75781e209131cea682df>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-04T04:57:16.897Z

Authors: Yiqin Lv, Qi Wang, Dong Liang, Zheng Xie

Published: 2024-10-30T08:07:43.000Z

## Abstract

Meta learning is a promising paradigm in the era of large models and task distributional robustness has become an indispensable consideration in real-world scenarios. Recent advances have examined the effectiveness of tail task risk minimization in fast adaptation robustness improvement \\citep{wang2023simple}. This work contributes to more theoretical investigations and practical enhancements in the field. Specifically, we reduce the distributionally robust strategy to a max-min optimization problem, constitute the Stackelberg equilibrium as the solution concept, and estimate the convergence rate. In the presence of tail risk, we further derive the generalization bound, establish connections with estimated quantiles, and practically improve the studied strategy. Accordingly, extensive evaluations demonstrate the significance of our proposal and its scalability to multimodal large models in boosting robustness.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Image Classification on mini-ImageNet (train)](<https://www.sota2.com/research/sota/image-classification-on-mini-imagenet-train.md>): Image Classification; mini-ImageNet (train); Average Score: 70.4.

- [System Identification on Pendulum (test)](<https://www.sota2.com/research/sota/system-identification-on-pendulum-test.md>): System Identification; Pendulum (test); Average MSE: 0.72.

- [Meta Reinforcement Learning on 2-D point robot navigation (meta-test)](<https://www.sota2.com/research/sota/meta-reinforcement-learning-on-2-d-point-robot-navigation-meta-test.md>): Meta-Reinforcement Learning; 2-D point robot navigation (meta-test); Average Return: -19.2.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2410.22788>)

[PDF](<https://arxiv.org/pdf/2410.22788.pdf>)

[Implementation](<https://github.com/lvyiqin/DRMAML>)