Multi-task Learning Curve Forecasting Across Hyperparameter Configurations and Datasets

Jawed, Shayan; Jomaa, Hadi S.; Schmidt-Thieme, Lars; Grabocka, Josif

doi:10.1007/978-3-030-86486-6_30

Cited by 4 publications

(2 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To the best of our knowledge, meta-learning for the task of eTSF as shown in Figure 1, is yet to be explored. However, the use of meta-learning methods for general time series forecasting have been explored previously in [4], [12]- [14]. The pioneering works [12], [13] showcased the applicability of well-established forecasting methods to the Zero-shot forecasting problem.…”

Section: * Equal Contributionmentioning

confidence: 99%

Forecasting Early with Meta Learning

Jawed

Madhusudhanan

Yalavarthi

et al. 2023

2023 International Joint Conference on Neural Networks (IJCNN)

Self Cite

View full text Add to dashboard Cite

In the early observation period of a time series, there might be only a few historic observations available to learn a model. However, in cases where an existing prior set of datasets is available, Meta learning methods can be applicable. In this paper, we devise a Meta learning method that exploits samples from additional datasets and learns to augment time series through adversarial learning as an auxiliary task for the target dataset. Our model (FEML), is equipped with a shared Convolutional backbone that learns features for varying length inputs from different datasets and has dataset specific heads to forecast for different output lengths. We show that FEML can meta learn across datasets and by additionally learning on adversarial generated samples as auxiliary samples for the target dataset, it can improve the forecasting performance compared to single task learning, and various solutions adapted from Joint learning, Multi-task learning and classic forecasting baselines.

show abstract

Section: * Equal Contributionmentioning

confidence: 99%

Forecasting Early with Meta Learning

Jawed

Madhusudhanan

Yalavarthi

et al. 2023

2023 International Joint Conference on Neural Networks (IJCNN)

Self Cite

View full text Add to dashboard Cite

show abstract

“…The current practice to find the correct trade-off weights is by hyperparameter tuning [22] using common search techniques. The complexity of such an approach can grow exponentially in time with each added task [12] and therefore does not scale well for a large number of tasks. Moreover, the trade-off weights in MTA-F are fixed throughout the training process.…”

Section: Introductionmentioning

confidence: 99%

Learning to Teach Fairness-aware Deep Multi-task Learning

Roy¹,

Ntoutsi²

2022

Preprint

View full text Add to dashboard Cite

Fairness-aware learning mainly focuses on single task learning (STL). The fairness implications of multi-task learning (MTL) have only recently been considered and a seminal approach has been proposed that considers the fairness-accuracy trade-off for each task and the performance trade-off among different tasks. Instead of a rigid fairnessaccuracy trade-off formulation, we propose a flexible approach that learns how to be fair in a MTL setting by selecting which objective (accuracy or fairness) to optimize at each step. We introduce the L2T-FMT algorithm that is a teacher-student network trained collaboratively; the student learns to solve the fair MTL problem while the teacher instructs the student to learn from either accuracy or fairness, depending on what is harder to learn for each task. Moreover, this dynamic selection of which objective to use at each step for each task reduces the number of trade-off weights from 2T to T , where T is the number of tasks. Our experiments on three real datasets show that L2T-FMT improves on both fairness (12-19%) and accuracy (up to 2%) over state-of-the-art approaches.

show abstract