Yan Ma

Yan Ma, Weicong Liang, Bohan Chen et al. · berkeley

Motivated by the remarkable achievements of DETR-based approaches on COCO object detection and segmentation benchmarks, recent endeavors have been directed towards elevating their performance through self-supervised pre-training of Transformers while preserving a frozen backbone. Noteworthy advancements in accuracy have been documented in certain studies. Our investigation delved deeply into a representative approach, DETReg, and its performance assessment in the context of emerging models like $\mathcal{H}$-Deformable-DETR. Regrettably, DETReg proves inadequate in enhancing the performance of robust DETR-based models under full data conditions. To dissect the underlying causes, we conduct extensive experiments on COCO and PASCAL VOC probing elements such as the selection of pre-training datasets and strategies for pre-training target generation. By contrast, we employ an optimized approach named Simple Self-training which leads to marked enhancements through the combination of an improved box predictor and the Objects$365$ benchmark. The culmination of these endeavors results in a remarkable AP score of $59.3\%$ on the COCO val set, outperforming $\mathcal{H}$-Deformable-DETR + Swin-L without pre-training by $1.4\%$. Moreover, a series of synthetic pre-training datasets, generated by merging contemporary image-to-text(LLaVA) and text-to-image (SDXL) models, significantly amplifies object detection capabilities.

Kang Xu, Yan Ma, Bingsheng Wei et al.

While Reinforcement Learning can achieve impressive results for complex tasks, the learned policies are generally prone to fail in downstream tasks with even minor model mismatch or unexpected perturbations. Recent works have demonstrated that a policy population with diverse behavior characteristics can generalize to downstream environments with various discrepancies. However, such policies might result in catastrophic damage during the deployment in practical scenarios like real-world systems due to the unrestricted behaviors of trained policies. Furthermore, training diverse policies without regulation of the behavior can result in inadequate feasible policies for extrapolating to a wide range of test conditions with dynamics shifts. In this work, we aim to train diverse policies under the regularization of the behavior patterns. We motivate our paradigm by observing the inverse dynamics in the environment with partial state information and propose Diversity in Regulation (DiR) training diverse policies with regulated behaviors to discover desired patterns that benefit the generalization. Considerable empirical results on various variations of different environments indicate that our method attains improvements over other diversity-driven counterparts.

Yuxiao Huang, Yan Ma

A key challenge in Machine Learning is class imbalance, where the sample size of some classes (majority classes) are much higher than that of the other classes (minority classes). If we were to train a classifier directly on imbalanced data, it is more likely for the classifier to predict a new sample as one of the majority classes. In the extreme case, the classifier could completely ignore the minority classes. This could have serious sociological implications in healthcare, as the minority classes are usually the disease classes (e.g., death or positive clinical test result). In this paper, we introduce a software that uses Generative Adversarial Networks to oversample the minority classes so as to improve downstream classification. To the best of our knowledge, this is the first tool that allows multi-class classification (where the target can have an arbitrary number of classes). The code of the tool is publicly available in our github repository (https://github.com/yuxiaohuang/research/tree/master/gwu/working/cigan/code).

Kang Xu, Yan Ma, Wei Li

Training a robust policy is critical for policy deployment in real-world systems or dealing with unknown dynamics mismatch in different dynamic systems. Domain Randomization~(DR) is a simple and elegant approach that trains a conservative policy to counter different dynamic systems without expert knowledge about the target system parameters. However, existing works reveal that the policy trained through DR tends to be over-conservative and performs poorly in target domains. Our key insight is that dynamic systems with different parameters provide different levels of difficulty for the policy, and the difficulty of behaving well in a system is constantly changing due to the evolution of the policy. If we can actively sample the systems with proper difficulty for the policy on the fly, it will stabilize the training process and prevent the policy from becoming over-conservative or over-optimistic. To operationalize this idea, we introduce Active Dynamics Preference~(ADP), which quantifies the informativeness and density of sampled system parameters. ADP actively selects system parameters with high informativeness and low density. We validate our approach in four robotic locomotion tasks with various discrepancies between the training and testing environments. Extensive results demonstrate that our approach has superior robustness for system inconsistency compared to several baselines.

Yan Ma, Tianxing Liu, Bingsheng Wei et al.

Evolutionary Algorithms (EAs) and Deep Reinforcement Learning (DRL) have recently been integrated to take the advantage of the both methods for better exploration and exploitation.The evolutionary part in these hybrid methods maintains a population of policy networks.However, existing methods focus on optimizing the parameters of policy network, which is usually high-dimensional and tricky for EA.In this paper, we shift the target of evolution from high-dimensional parameter space to low-dimensional action space.We propose Evolutionary Action Selection-Twin Delayed Deep Deterministic Policy Gradient (EAS-TD3), a novel hybrid method of EA and DRL.In EAS, we focus on optimizing the action chosen by the policy network and attempt to obtain high-quality actions to promote policy learning through an evolutionary algorithm. We conduct several experiments on challenging continuous control tasks.The result shows that EAS-TD3 shows superior performance over other state-of-art methods.

Zhengzheng Tu, Yan Ma, Zhun Li et al.

Salient object detection in complex scenes and environments is a challenging research topic. Most works focus on RGB-based salient object detection, which limits its performance of real-life applications when confronted with adverse conditions such as dark environments and complex backgrounds. Taking advantage of RGB and thermal infrared images becomes a new research direction for detecting salient object in complex scenes recently, as thermal infrared spectrum imaging provides the complementary information and has been applied to many computer vision tasks. However, current research for RGBT salient object detection is limited by the lack of a large-scale dataset and comprehensive benchmark. This work contributes such a RGBT image dataset named VT5000, including 5000 spatially aligned RGBT image pairs with ground truth annotations. VT5000 has 11 challenges collected in different scenes and environments for exploring the robustness of algorithms. With this dataset, we propose a powerful baseline approach, which extracts multi-level features within each modality and aggregates these features of all modalities with the attention mechanism, for accurate RGBT salient object detection. Extensive experiments show that the proposed baseline approach outperforms the state-of-the-art methods on VT5000 dataset and other two public datasets. In addition, we carry out a comprehensive analysis of different algorithms of RGBT salient object detection on VT5000 dataset, and then make several valuable conclusions and provide some potential research directions for RGBT salient object detection.