Hao Sun

Yaming Yang, Dilxat Muhtar, Yelong Shen et al.

Parameter-efficient fine-tuning (PEFT) has been widely employed for domain adaptation, with LoRA being one of the most prominent methods due to its simplicity and effectiveness. However, in multi-task learning (MTL) scenarios, LoRA tends to obscure the distinction between tasks by projecting sparse high-dimensional features from different tasks into the same dense low-dimensional intrinsic space. This leads to task interference and suboptimal performance for LoRA and its variants. To tackle this challenge, we propose MTL-LoRA, which retains the advantages of low-rank adaptation while significantly enhancing MTL capabilities. MTL-LoRA augments LoRA by incorporating additional task-adaptive parameters that differentiate task-specific information and capture shared knowledge across various tasks within low-dimensional spaces. This approach enables pre-trained models to jointly adapt to different target domains with a limited number of trainable parameters. Comprehensive experimental results, including evaluations on public academic benchmarks for natural language understanding, commonsense reasoning, and image-text understanding, as well as real-world industrial text Ads relevance datasets, demonstrate that MTL-LoRA outperforms LoRA and its various variants with comparable or even fewer learnable parameters in MTL setting.

Dilxat Muhtar, Yelong Shen, Yaming Yang et al.

In-context learning (ICL) allows large language models (LLMs) to adapt to new tasks directly from the given demonstrations without requiring gradient updates. While recent advances have expanded context windows to accommodate more demonstrations, this approach increases inference costs without necessarily improving performance. To mitigate these issues, We propose StreamAdapter, a novel approach that directly updates model parameters from context at test time, eliminating the need for explicit in-context demonstrations. StreamAdapter employs context mapping and weight absorption mechanisms to dynamically transform ICL demonstrations into parameter updates with minimal additional parameters. By reducing reliance on numerous in-context examples, StreamAdapter significantly reduce inference costs and allows for efficient inference with constant time complexity, regardless of demonstration count. Extensive experiments across diverse tasks and model architectures demonstrate that StreamAdapter achieves comparable or superior adaptation capability to ICL while requiring significantly fewer demonstrations. The superior task adaptation and context encoding capabilities of StreamAdapter on both language understanding and generation tasks provides a new perspective for adapting LLMs at test time using context, allowing for more efficient adaptation across scenarios and more cost-effective inference

Hao Sun, Jiayi Wu, Hengyi Cai et al.

Recent advancements in large language models (LLMs) have been remarkable. Users face a choice between using cloud-based LLMs for generation quality and deploying local-based LLMs for lower computational cost. The former option is typically costly and inefficient, while the latter usually fails to deliver satisfactory performance for reasoning steps requiring deliberate thought processes. In this work, we propose a novel LLM utilization paradigm that facilitates the collaborative operation of large cloud-based LLMs and smaller local-deployed LLMs. Our framework comprises two primary modules: the local agent instantiated with a relatively smaller LLM, handling less complex reasoning steps, and the cloud agent equipped with a larger LLM, managing more intricate reasoning steps. This collaborative processing is enabled through an adaptive mechanism where the local agent introspectively identifies errors and proactively seeks assistance from the cloud agent, thereby effectively integrating the strengths of both locally-deployed and cloud-based LLMs, resulting in significant enhancements in task completion performance and efficiency. We evaluate AdaSwitch across 7 benchmarks, ranging from mathematical reasoning and complex question answering, using various types of LLMs to instantiate the local and cloud agents. The empirical results show that AdaSwitch effectively improves the performance of the local agent, and sometimes achieves competitive results compared to the cloud agent while utilizing much less computational overhead.

Xu Yang, Qi Zhang, Shuming Jiang et al.

With the rapid advancement of generative AI, synthetic content across images, videos, and audio has become increasingly realistic, amplifying the risk of misinformation. Existing detection approaches predominantly focus on binary classification while lacking detailed and interpretable explanations of forgeries, which limits their applicability in safety-critical scenarios. Moreover, current methods often treat each modality separately, without a unified benchmark for cross-modal forgery detection and interpretation. To address these challenges, we introduce METER, a unified, multi-modal benchmark for interpretable forgery detection spanning images, videos, audio, and audio-visual content. Our dataset comprises four tracks, each requiring not only real-vs-fake classification but also evidence-chain-based explanations, including spatio-temporal localization, textual rationales, and forgery type tracing. Compared to prior benchmarks, METER offers broader modality coverage and richer interpretability metrics such as spatial/temporal IoU, multi-class tracing, and evidence consistency. We further propose a human-aligned, three-stage Chain-of-Thought (CoT) training strategy combining SFT, DPO, and a novel GRPO stage that integrates a human-aligned evaluator with CoT reasoning. We hope METER will serve as a standardized foundation for advancing generalizable and interpretable forgery detection in the era of generative media.

Zhixiao Wu, Yao Lu, Jie Wen et al.

Poison-only Clean-label Backdoor Attacks aim to covertly inject attacker-desired behavior into DNNs by merely poisoning the dataset without changing the labels. To effectively implant a backdoor, multiple \textbf{triggers} are proposed for various attack requirements of Attack Success Rate (ASR) and stealthiness. Additionally, sample selection enhances clean-label backdoor attacks' ASR by meticulously selecting ``hard'' samples instead of random samples to poison. Current methods 1) usually handle the sample selection and triggers in isolation, leading to severely limited improvements on both ASR and stealthiness. Consequently, attacks exhibit unsatisfactory performance on evaluation metrics when converted to PCBAs via a mere stacking of methods. Therefore, we seek to explore the bidirectional collaborative relations between the sample selection and triggers to address the above dilemma. 2) Since the strong specificity within triggers, the simple combination of sample selection and triggers fails to substantially enhance both evaluation metrics, with generalization preserved among various attacks. Therefore, we seek to propose a set of components to significantly improve both stealthiness and ASR based on the commonalities of attacks. Specifically, Component A ascertains two critical selection factors, and then makes them an appropriate combination based on the trigger scale to select more reasonable ``hard'' samples for improving ASR. Component B is proposed to select samples with similarities to relevant trigger implanted samples to promote stealthiness. Component C reassigns trigger poisoning intensity on RGB colors through distinct sensitivity of the human visual system to RGB for higher ASR, with stealthiness ensured by sample selection, including Component B. Furthermore, all components can be strategically integrated into diverse PCBAs.

Hao Sun, Ziping Xu, Meng Fang et al.

Safe exploration is crucial for the real-world application of reinforcement learning (RL). Previous works consider the safe exploration problem as Constrained Markov Decision Process (CMDP), where the policies are being optimized under constraints. However, when encountering any potential dangers, human tends to stop immediately and rarely learns to behave safely in danger. Motivated by human learning, we introduce a new approach to address safe RL problems under the framework of Early Terminated MDP (ET-MDP). We first define the ET-MDP as an unconstrained MDP with the same optimal value function as its corresponding CMDP. An off-policy algorithm based on context models is then proposed to solve the ET-MDP, which thereby solves the corresponding CMDP with better asymptotic performance and improved learning efficiency. Experiments on various CMDP tasks show a substantial improvement over previous methods that directly solve CMDP.

Zhenghao Peng, Hao Sun, Bolei Zhou

Conventional Reinforcement Learning (RL) algorithms usually have one single agent learning to solve the task independently. As a result, the agent can only explore a limited part of the state-action space while the learned behavior is highly correlated to the agent's previous experience, making the training prone to a local minimum. In this work, we empower RL with the capability of teamwork and propose a novel non-local policy optimization framework called Diversity-regularized Collaborative Exploration (DiCE). DiCE utilizes a group of heterogeneous agents to explore the environment simultaneously and share the collected experiences. A regularization mechanism is further designed to maintain the diversity of the team and modulate the exploration. We implement the framework in both on-policy and off-policy settings and the experimental results show that DiCE can achieve substantial improvement over the baselines in the MuJoCo locomotion tasks.

Hao Sun, Xinyu Pan, Bo Dai et al.

Solving the Goal-Conditioned Reward Sparse (GCRS) task is a challenging reinforcement learning problem due to the sparsity of reward signals. In this work, we propose a new formulation of GCRS tasks from the perspective of the drifted random walk on the state space, and design a novel method called Evolutionary Stochastic Policy Distillation (ESPD) to solve them based on the insight of reducing the First Hitting Time of the stochastic process. As a self-imitate approach, ESPD enables a target policy to learn from a series of its stochastic variants through the technique of policy distillation (PD). The learning mechanism of ESPD can be considered as an Evolution Strategy (ES) that applies perturbations upon policy directly on the action space, with a SELECT function to check the superiority of stochastic variants and then use PD to update the policy. The experiments based on the MuJoCo robotics control suite show the high learning efficiency of the proposed method.

Hao Sun, Jiadong Guo, Edward J. Kim et al.

The increasing amount of data in astronomy provides great challenges for machine learning research. Previously, supervised learning methods achieved satisfactory recognition accuracy for the star-galaxy classification task, based on manually labeled data set. In this work, we propose a novel unsupervised approach for the star-galaxy recognition task, namely Cascade Variational Auto-Encoder (CasVAE). Our empirical results show our method outperforms the baseline model in both accuracy and stability.

Yang Li, Jianhe Yuan, Zhiqun Zhao et al.

In this work, we develop a joint sample discovery and iterative model evolution method for semi-supervised learning on very small labeled training sets. We propose a master-teacher-student model framework to provide multi-layer guidance during the model evolution process with multiple iterations and generations. The teacher model is constructed by performing an exponential moving average of the student models obtained from past training steps. The master network combines the knowledge of the student and teacher models with additional access to newly discovered samples. The master and teacher models are then used to guide the training of the student network by enforcing the consistence between their predictions of unlabeled samples and evolve all models when more and more samples are discovered. Our extensive experiments demonstrate that the discovering confident samples from the unlabeled dataset, once coupled with the above master-teacher-student network evolution, can significantly improve the overall semi-supervised learning performance. For example, on the CIFAR-10 dataset, with a very small set of 250 labeled samples, our method achieves an error rate of 11.81 %, more than 38 % lower than the state-of-the-art method Mean-Teacher (49.91 %).

Qianggang Ding, Sifan Wu, Hao Sun et al.

Recently, a variety of regularization techniques have been widely applied in deep neural networks, such as dropout, batch normalization, data augmentation, and so on. These methods mainly focus on the regularization of weight parameters to prevent overfitting effectively. In addition, label regularization techniques such as label smoothing and label disturbance have also been proposed with the motivation of adding a stochastic perturbation to labels. In this paper, we propose a novel adaptive label regularization method, which enables the neural network to learn from the erroneous experience and update the optimal label representation online. On the other hand, compared with knowledge distillation, which learns the correlation of categories using teacher network, our proposed method requires only a minuscule increase in parameters without cumbersome teacher network. Furthermore, we evaluate our method on CIFAR-10/CIFAR-100/ImageNet datasets for image recognition tasks and AGNews/Yahoo/Yelp-Full datasets for text classification tasks. The empirical results show significant improvement under all experimental settings.

Xue Yang, Hao Sun, Xian Sun et al.

Ship detection is of great importance and full of challenges in the field of remote sensing. The complexity of application scenarios, the redundancy of detection region, and the difficulty of dense ship detection are all the main obstacles that limit the successful operation of traditional methods in ship detection. In this paper, we propose a brand new detection model based on multitask rotational region convolutional neural network to solve the problems above. This model is mainly consist of five consecutive parts: Dense Feature Pyramid Network (DFPN), adaptive region of interest (ROI) Align, rotational bounding box regression, prow direction prediction and rotational nonmaximum suppression (R-NMS). First of all, the low-level location information and high-level semantic information are fully utilized through multiscale feature networks. Then, we design Adaptive ROI Align to obtain high quality proposals which remain complete spatial and semantic information. Unlike most previous approaches, the prediction obtained by our method is the minimum bounding rectangle of the object with less redundant regions. Therefore, rotational region detection framework is more suitable to detect the dense object than traditional detection model. Additionally, we can find the berthing and sailing direction of ship through prediction. A detailed evaluation based on SRSS for rotation detection shows that our detection method has a competitive performance.

Xue Yang, Hao Sun, Kun Fu et al.

Ship detection has been playing a significant role in the field of remote sensing for a long time but it is still full of challenges. The main limitations of traditional ship detection methods usually lie in the complexity of application scenarios, the difficulty of intensive object detection and the redundancy of detection region. In order to solve such problems above, we propose a framework called Rotation Dense Feature Pyramid Networks (R-DFPN) which can effectively detect ship in different scenes including ocean and port. Specifically, we put forward the Dense Feature Pyramid Network (DFPN), which is aimed at solving the problem resulted from the narrow width of the ship. Compared with previous multi-scale detectors such as Feature Pyramid Network (FPN), DFPN builds the high-level semantic feature-maps for all scales by means of dense connections, through which enhances the feature propagation and encourages the feature reuse. Additionally, in the case of ship rotation and dense arrangement, we design a rotation anchor strategy to predict the minimum circumscribed rectangle of the object so as to reduce the redundant detection region and improve the recall. Furthermore, we also propose multi-scale ROI Align for the purpose of maintaining the completeness of semantic and spatial information. Experiments based on remote sensing images from Google Earth for ship detection show that our detection method based on R-DFPN representation has a state-of-the-art performance.

Liangbo He, Hao Sun

In traditional neural networks for image processing, the inputs of the neural networks should be the same size such as 224*224*3. But how can we train the neural net model with different input size? A common way to do is image deformation which accompany a problem of information loss (e.g. image crop or wrap). Sequence model(RNN, LSTM, etc.) can accept different size of input like text and audio. But one disadvantage for sequence model is that the previous information will become more fragmentary during the transfer in time step, it will make the network hard to train especially for long sequential data. In this paper we propose a new network structure called Attention Incorporate Network(AIN). It solve the problem of different size of inputs including: images, text, audio, and extract the key features of the inputs by attention mechanism, pay different attention depends on the importance of the features not rely on the data size. Experimentally, AIN achieve a higher accuracy, better convergence comparing to the same size of other network structure

Zhi Zhang, Guanghan Ning, Yigang Cen et al.

The inference structures and computational complexity of existing deep neural networks, once trained, are fixed and remain the same for all test images. However, in practice, it is highly desirable to establish a progressive structure for deep neural networks which is able to adapt its inference process and complexity for images with different visual recognition complexity. In this work, we develop a multi-stage progressive structure with integrated confidence analysis and decision policy learning for deep neural networks. This new framework consists of a set of network units to be activated in a sequential manner with progressively increased complexity and visual recognition power. Our extensive experimental results on the CIFAR-10 and ImageNet datasets demonstrate that the proposed progressive deep neural network is able to obtain more than 10 fold complexity scalability while achieving the state-of-the-art performance using a single network model satisfying different complexity-accuracy requirements.