Xiaofeng Wu

Guoqing Wang, Sunhao Dai, Guangze Ye et al.

Large language model (LLM)-based agents are increasingly trained with reinforcement learning (RL) to enhance their ability to interact with external environments through tool use, particularly in search-based settings that require multi-turn reasoning and knowledge acquisition. However, existing approaches typically rely on outcome-based rewards that are only provided at the final answer. This reward sparsity becomes particularly problematic in multi-turn settings, where long trajectories exacerbate two critical issues: (i) advantage collapse, where all rollouts receive identical rewards and provide no useful learning signals, and (ii) lack of fine-grained credit assignment, where dependencies between turns are obscured, especially in long-horizon tasks. In this paper, we propose Information Gain-based Policy Optimization (IGPO), a simple yet effective RL framework that provides dense and intrinsic supervision for multi-turn agent training. IGPO models each interaction turn as an incremental process of acquiring information about the ground truth, and defines turn-level rewards as the marginal increase in the policy's probability of producing the correct answer. Unlike prior process-level reward approaches that depend on external reward models or costly Monte Carlo estimation, IGPO derives intrinsic rewards directly from the model's own belief updates. These intrinsic turn-level rewards are combined with outcome-level supervision to form dense reward trajectories. Extensive experiments on both in-domain and out-of-domain benchmarks demonstrate that IGPO consistently outperforms strong baselines in multi-turn scenarios, achieving higher accuracy and improved sample efficiency.

Yubo Yang, Tao Yang, Xiaofeng Wu et al.

The rapid development of Unmanned aerial vehicles (UAVs) technology has spawned a wide variety of applications, such as emergency communications, regional surveillance, and disaster relief. Due to their limited battery capacity and processing power, multiple UAVs are often required for complex tasks. In such cases, a control center is crucial for coordinating their activities, which fits well with the federated learning (FL) framework. However, conventional FL approaches often focus on a single task, ignoring the potential of training multiple related tasks simultaneously. In this paper, we propose a UAV-assisted multi-task federated learning scheme, in which data collected by multiple UAVs can be used to train multiple related tasks concurrently. The scheme facilitates the training process by sharing feature extractors across related tasks and introduces a task attention mechanism to balance task performance and encourage knowledge sharing. To provide an analytical description of training performance, the convergence analysis of the proposed scheme is performed. Additionally, the optimal bandwidth allocation for UAVs under limited bandwidth conditions is derived to minimize communication time. Meanwhile, a UAV-EV association strategy based on coalition formation game is proposed. Simulation results validate the effectiveness of the proposed scheme in enhancing multi-task performance and training speed.

Haolin Fei, Xiaofeng Wu, Chunbo Luo

Air quality is closely related to public health. Health issues such as cardiovascular diseases and respiratory diseases, may have connection with long exposure to highly polluted environment. Therefore, accurate air quality forecasts are extremely important to those who are vulnerable. To estimate the variation of several air pollution concentrations, previous researchers used various approaches, such as the Community Multiscale Air Quality model (CMAQ) or neural networks. Although CMAQ model considers a coverage of the historic air pollution data and meteorological variables, extra bias is introduced due to additional adjustment. In this paper, a combination of model-based strategy and data-driven method namely the physical-temporal collection(PTC) model is proposed, aiming to fix the systematic error that traditional models deliver. In the data-driven part, the first components are the temporal pattern and the weather pattern to measure important features that contribute to the prediction performance. The less relevant input variables will be removed to eliminate negative weights in network training. Then, we deploy a long-short-term-memory (LSTM) to fetch the preliminary results, which will be further corrected by a neural network (NN) involving the meteorological index as well as other pollutants concentrations. The data-set we applied for forecasting is from January 1st, 2016 to December 31st, 2016. According to the results, our PTC achieves an excellent performance compared with the baseline model (CMAQ prediction, GRU, DNN and etc.). This joint model-based data-driven method for air quality prediction can be easily deployed on stations without extra adjustment, providing results with high-time-resolution information for vulnerable members to prevent heavy air pollution ahead.

Hui Yu, Xiaofeng Wu, Wenbin Jiang et al.

The widely-used automatic evaluation metrics cannot adequately reflect the fluency of the translations. The n-gram-based metrics, like BLEU, limit the maximum length of matched fragments to n and cannot catch the matched fragments longer than n, so they can only reflect the fluency indirectly. METEOR, which is not limited by n-gram, uses the number of matched chunks but it does not consider the length of each chunk. In this paper, we propose an entropy-based method, which can sufficiently reflect the fluency of translations through the distribution of matched words. This method can easily combine with the widely-used automatic evaluation metrics to improve the evaluation of fluency. Experiments show that the correlations of BLEU and METEOR are improved on sentence level after combining with the entropy-based method on WMT 2010 and WMT 2012.

Hui Yu, Xiaofeng Wu, Wenbin Jiang et al.

Most of the syntax-based metrics obtain the similarity by comparing the sub-structures extracted from the trees of hypothesis and reference. These sub-structures are defined by human and can't express all the information in the trees because of the limited length of sub-structures. In addition, the overlapped parts between these sub-structures are computed repeatedly. To avoid these problems, we propose a novel automatic evaluation metric based on dependency parsing model, with no need to define sub-structures by human. First, we train a dependency parsing model by the reference dependency tree. Then we generate the hypothesis dependency tree and the corresponding probability by the dependency parsing model. The quality of the hypothesis can be judged by this probability. In order to obtain the lexicon similarity, we also introduce the unigram F-score to the new metric. Experiment results show that the new metric gets the state-of-the-art performance on system level, and is comparable with METEOR on sentence level.