Product retrieval is the backbone of e-commerce search: for each user query, it identifies a high-recall candidate set from billions of items, laying the foundation for high-quality ranking and user experience. Despite extensive optimization for mainstream queries, existing systems still struggle with long-tail queries, especially knowledge-intensive ones. These queries exhibit diverse linguistic patterns, often lack explicit purchase intent, and require domain-specific knowledge reasoning for accurate interpretation. They also suffer from a shortage of reliable behavioral logs, which makes such queries a persistent challenge for retrieval optimization.

In the realm of e-commerce search, semantic matching has consistently been a core issue, as it directly affects user experience and company revenue. However, users' queries often fail to effectively retrieve relevant products due to discrepancies between the user's expression habits and product names written by merchants. Even existing large language model (LLM) based query rewriting methods can bridge the semantic gap for most queries, they are still ineffective for long-tail queries with complicated semantic. In this paper, we propose Complicated Semantic Alignment Query Rewrite(CSA-QR) framework, which mitigates the semantic differences in long-tail queries with complicated semantics. CSA-QR comprises three stages: high-quality supervised fine-tuning (SFT) dataset generation, multi-dimensional alignment dataset generation, and binary feedback Proximal Policy Optimization (PPO) for reinforcement alignment. Initially, we utilize general large language models to generate rewrite candidates, followed by manual annotation to discriminate the candidates, then use the retrieval augmentation generation (RAG) based on existing annotations to produce a higher quality SFT dataset. Subsequently, we decouple the feedback data into user semantic consistency and merchant expression consistency dimensions to collect multi-dimensional alignment data. Finally, we introduce a binary feedback method to train the reward model, enabling it to better guide alignment training within our context. We also identify a set of more appropriate reward model evaluation metrics to guide our iterations. Offline experiments demonstrate the effectiveness of this method in improving retrieval performance. Online A/B tests reveal that our method significantly boosts critical metrics such as product click-through rate (CTR), gross merchandise volume (GMV) and number of transaction (#Trans) for long-tail complicated queries. CSA-QR has been deployed on Taobao, one of China's most popular online shopping platforms, since September 2024.

Large language models (LLMs) have garnered significant attention for their remarkable capabilities across various domains, whose vast parameter scales present challenges for practical deployment. Structured pruning is an effective method to balance model performance with efficiency, but performance restoration under computational resource constraints is a principal challenge in pruning LLMs. Therefore, we present a low-cost and fast structured pruning method for LLMs named SlimGPT based on the Optimal Brain Surgeon framework. We propose Batched Greedy Pruning for rapid and near-optimal pruning, which enhances the accuracy of head-wise pruning error estimation through grouped Cholesky decomposition and improves the pruning efficiency of FFN via Dynamic Group Size, thereby achieving approximate local optimal pruning results within one hour. Besides, we explore the limitations of layer-wise pruning from the perspective of error accumulation and propose Incremental Pruning Ratio, a non-uniform pruning strategy to reduce performance degradation. Experimental results on the LLaMA benchmark show that SlimGPT outperforms other methods and achieves state-of-the-art results.

As a popular robot path planning algorithm, RRT (Rapid-exploring Random Tree) and various RRT-based extensions have achieved remarkable results. However, existing algorithms rely too much on randomness and often do not make use of known map information, this makes the growth of trees too blind. To this end, we introduce Bidirectional Homotopy-Guided RRT (BH-RRT) that combines Bidirectional RRT (Bi-RRT) with information obtained from obstacle contours. Compared with the previous methods, BH-RRT can reflect most of the map's information with a small number of feature points, and then form a set of points. This set can provide a better local goal point in each iteration, so that the tree can grow in a more favorable direction, rather than only being affected by the goal point. Experimental results show that BH-RRT outperforms RRT, HRRT and Bi-RRT in the success rate within a limited time.

This paper introduces a fishery ecology monitoring system for cultivation pools, and proposes a new stereo keypoint detection method followed by curve fitting analysis to estimate the fish posture and length. The system, which can be employed for aquaculture monitoring, is featured by its exploitation of underwater visual intelligence and deep neural-network architecture. As input, stereo image pairs are obtained by underwater binocular camera. A deep neural-network under Faster R-CNN architecture is built to detect fish from the stereo image inputs. Another network under Stacked Hourglass architecture is constructed to detect specific keypoints of each fish. For ecology monitoring, detected keypoints are used in the estimation of the fishes' posture and length. Unlike other size estimation methods which also apply a binocular camera, our method naturally bypasses the pixel-wise matching difficulty in global stereo matching algorithms. Experiment shows that our system is applicable for online fish ecology monitoring, with efficient and accurate estimation performance.