Author name cluster

Jialu Gao

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

3 papers

1 author row

EAAI Journal 2026 Journal Article

A surrogate model for single-stage synchronous induction coilgun based on scientific machine learning

Jialu Gao
Gangquan Si
Xikui Ma
Jiawei Wang
Yating Wang

The synchronous induction coilgun is a representative electromechanical coupled system whose numerical modeling usually relies on the finite element method (FEM) and the current filament method (CFM). However, these physics-based approaches are often computationally intensive and time-consuming in design and optimization processes. This study aims to develop a data-efficient operator-learning framework, termed Gated Recurrent Unit and Fully Connected Operator Network (G&F-ONet), for predicting the velocity trajectory of a single-stage induction coilgun. Built upon the Deep Operator Network (DeepONet) architecture, the proposed model embeds physical prior knowledge by introducing a recurrent branch whose structure reflects the evolution pattern of geometric parameters during launch. The dual-branch vanilla Multiple-Input Operator Network (MIONet) is used as the baseline, and the relative L 2 error serves as the primary evaluation metric. Experimental results show that, using a dataset generated through Latin Hypercube Sampling (LHS) from a CFM-based model, G&F-ONet reduces the average relative L 2 error by 43. 1% compared with the baseline model and exhibits superior data efficiency, robustness, and generalization capability. All experiments are performed on an NVIDIA RTX 2060 Graphics Processing Unit (GPU) with an Intel i5 Central Processing Unit (CPU). The results demonstrate that the proposed approach, which introduces physical priors through inductive bias, can effectively learn the complex mapping between coilgun parameters and launch velocity, achieving fast, reliable, and physically interpretable predictions. Overall, the proposed framework provides an efficient surrogate model for the design and optimization of synchronous induction coilguns.

Details DOI

NeurIPS Conference 2024 Conference Paper

Visual Data Diagnosis and Debiasing with Concept Graphs

Rwiddhi Chakraborty
Yinong Wang
Jialu Gao
Runkai Zheng
Cheng Zhang
Fernando De la Torre

The widespread success of deep learning models today is owed to the curation of extensive datasets significant in size and complexity. However, such models frequently pick up inherent biases in the data during the training process, leading to unreliable predictions. Diagnosing and debiasing datasets is thus a necessity to ensure reliable model performance. In this paper, we present ConBias, a novel framework for diagnosing and mitigating Concept co-occurrence Biases in visual datasets. ConBias represents visual datasets as knowledge graphs of concepts, enabling meticulous analysis of spurious concept co-occurrences to uncover concept imbalances across the whole dataset. Moreover, we show that by employing a novel clique-based concept balancing strategy, we can mitigate these imbalances, leading to enhanced performance on downstream tasks. Extensive experiments show that data augmentation based on a balanced concept distribution augmented by ConBias improves generalization performance across multiple datasets compared to state-of-the-art methods.

PDF Details DOI

NeurIPS Conference 2023 Conference Paper

Can Pre-Trained Text-to-Image Models Generate Visual Goals for Reinforcement Learning?

Jialu Gao
Kaizhe Hu
Guowei Xu
Huazhe Xu

Pre-trained text-to-image generative models can produce diverse, semantically rich, and realistic images from natural language descriptions. Compared with language, images usually convey information with more details and less ambiguity. In this study, we propose Learning from the Void (LfVoid), a method that leverages the power of pre-trained text-to-image models and advanced image editing techniques to guide robot learning. Given natural language instructions, LfVoid can edit the original observations to obtain goal images, such as "wiping" a stain off a table. Subsequently, LfVoid trains an ensembled goal discriminator on the generated image to provide reward signals for a reinforcement learning agent, guiding it to achieve the goal. The ability of LfVoid to learn with zero in-domain training on expert demonstrations or true goal observations (the void) is attributed to the utilization of knowledge from web-scale generative models. We evaluate LfVoid across three simulated tasks and validate its feasibility in the corresponding real-world scenarios. In addition, we offer insights into the key considerations for the effective integration of visual generative models into robot learning workflows. We posit that our work represents an initial step towards the broader application of pre-trained visual generative models in the robotics field. Our project page: https: //lfvoid-rl. github. io/.

PDF Details