Learning to solve large-scale Mixed Integer Program (MIP) problems is an emerging research topic, and policy learning-based Large Neighborhood Search (LNS) has been a popular paradigm. However, the explored space of LNS policy is often limited even in the training phase, making the learned policy sometimes wrongly fix some potentially important variables early in the search, leading to local optimum in some cases. Moreover, many methods only assume binary variables to deal with. We present a practical approach, termed Binarized-Tightening Branch-and-Search for Large Neighborhood Search (BTBS-LNS). It comprises three key techniques: 1) the ``Binarized Tightening" technique for integer variables to handle their wide range by binary encoding and bound tightening; 2) an attention-based tripartite graph to capture global correlations among variables and constraints for an MIP instance; 3) an extra branching network as a global view, to identify and optimize wrongly-fixed backdoor variables at each search step. Experiments show its superior performance over the open-source solver SCIP and LNS baselines. Moreover, it performs competitively with, and sometimes better than the commercial solver Gurobi (v9.5.0), especially on the MIPLIB2017 benchmark chosen by Hans Mittelmann, where our method can deliver 10\% better primal gaps compared with Gurobi in a 300s cut-off time.

Foundation Models (FMs) have demonstrated remarkable success in fields like computer vision and natural language processing, yet their application to combinatorial optimization remains underexplored. Optimization problems, often modeled as graphs, pose unique challenges due to their diverse structures, varying distributions, and NP-hard complexity. To address these challenges, we propose OPTFM, the first graph foundation model for general combinatorial optimization. OPTFM introduces a scalable multi-view graph transformer with hybrid self-attention and cross-attention to model large-scale heterogeneous graphs in $O(N)$ time complexity while maintaining semantic consistency throughout the attention computation. A Dual-level pre-training framework integrates node-level graph reconstruction and instance-level contrastive learning, enabling robust and adaptable representations at multiple levels. Experimental results across diverse optimization tasks show that models trained on OPTFM embeddings without fine-tuning consistently outperform task-specific approaches, establishing a new benchmark for solving combinatorial optimization problems.

Branch-and-bound (B\&B) has long been favored for tackling complex Mixed Integer Programming (MIP) problems, where the choice of branching strategy plays a pivotal role. Recently, Imitation Learning (IL)-based policies have emerged as potent alternatives to traditional rule-based approaches. However, it is nontrivial to acquire high-quality training samples, and IL often converges to suboptimal variable choices for branching, restricting the overall performance. In response to these challenges, we propose a novel hybrid online and offline reinforcement learning (RL) approach to enhance the branching policy by cost-effective training sample augmentation. In the online phase, we train an online RL agent to dynamically decide the sample generation processes, drawing from either the learning-based policy or the expert policy. The objective is to strike a balance between exploration and exploitation of the sample generation process. In the offline phase, a value function is trained to fit each decision's cumulative reward and filter the samples with high cumulative returns. This dual-purpose function not only reduces training complexity but also enhances the quality of the samples. To assess the efficacy of our data augmentation mechanism, we conduct comprehensive evaluations across a range of MIP problems. The results consistently show that it excels in making superior branching decisions compared to state-of-the-art learning-based models and the open-source solver SCIP. Notably, it even often outperforms Gurobi.

The field of scene text detection has gained increasing attention in recent years. However, it faces the challenges of high computational overhead. This paper proposes a Lightweight Multi-Step Feature Fusion Network (LMFFNet), an enhanced model designed to reduce computational burden in scene text detection tasks. Based on the lightweight MobileNetV3 backbone network, the proposed network combines the Triplet Attention Feature Enhancement Module (TFEM) and the Dynamic Feature Fusion Module (DFFM) to achieve efficient scene text detection. The TFEM expands the receptive field of the model and better utilizes features of various scales. The DFFM, which incorporates Convolutional Block Attention Module (CBAM), adaptively fuses features of different scales to improve the performance and robustness of the model. In the loss function, Dice Loss is used to address the imbalance between positive and negative samples. The proposed LMFFNet achieves a detection accuracy of 90.7% on the ICDAR2015 dataset, with a 91.7% reduction in parameter count and a 70.0% reduction in overall model overhead compared to the baseline model DBNet (ResNet-50).The effectiveness of the lightweight scene text detector is verified.

The Convolution Neural Network (CNN) has achieved great performance in image classification, partially due to the deeper and deeper structure. While its complexity brings more challenge to the practical application. So we can design a more efficient loss function to get better results without a more complex network. In this paper, we proposed a weighted softmax loss function called confusion weighted loss to learn the relationship among the confusing categories. Firstly, we generate a similarity matrix based on the confusion matrix to illustrate the relationship among the categories. Then, we propose a clustering algorithm to find out the confusing categories. Finally, to learn more information among them, we design a weighted matrix used in our loss function. Experiments on different datasets demonstrate the effectiveness of our method.