REVIEW 5 cited by
Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Models
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Massively multilingual models subsuming tens or even hundreds of languages pose great challenges to multi-task optimization. While it is a common practice to apply a language-agnostic procedure optimizing a joint multilingual task objective, how to properly characterize and take advantage of its underlying problem structure for improving optimization efficiency remains under-explored. In this paper, we attempt to peek into the black-box of multilingual optimization through the lens of loss function geometry. We find that gradient similarity measured along the optimization trajectory is an important signal, which correlates well with not only language proximity but also the overall model performance. Such observation helps us to identify a critical limitation of existing gradient-based multi-task learning methods, and thus we derive a simple and scalable optimization procedure, named Gradient Vaccine, which encourages more geometrically aligned parameter updates for close tasks. Empirically, our method obtains significant model performance gains on multilingual machine translation and XTREME benchmark tasks for multilingual language models. Our work reveals the importance of properly measuring and utilizing language proximity in multilingual optimization, and has broader implications for multi-task learning beyond multilingual modeling.
Forward citations
Cited by 5 Pith papers
-
Exploring Line Bundle Standard Models with Transformers
A Transformer trained by reinforcement learning generates heterotic line-bundle sums that satisfy anomaly-cancellation, stability, and chirality constraints, and its policy transfers usefully across Calabi-Yau geometries.
-
The Parts Are Greater Than the Sum: Automated Task Sequencing for Efficient Training of Multi-Policy LLMs
Automatically grouping and sequencing tasks into multiple QLoRA adapters improves continual fine-tuning performance over a single shared adapter at matched trainable capacity.
-
Multi-Task GRPO: Reliable LLM Reasoning Across Tasks
MT-GRPO reweights tasks by reward and improvement and enforces those weights after zero-gradient filtering, improving worst-task accuracy by 6–28% over GRPO/DAPO baselines on 3- and 9-task setups.
-
FastCAR: Fast Classification And Regression for Task Consolidation in Multi-Task Learning to Model a Continuous Property Variable of Detected Object Class
A single regression network trained on class-shifted hardness labels can jointly classify steel microstructures and predict hardness, outperforming multi-task baselines on the authors' own dataset.
-
DeepChest: Dynamic Gradient-Free Task Weighting for Effective Multi-Task Learning in Chest X-ray Classification
DeepChest weights each chest X-ray pathology task by comparing its current training accuracy to the average, boosting weak tasks and shrinking strong ones.
Discussion (0). Sign in to comment.