Contour accuracy is crucial for machining precision in five-axis computer numerical control (CNC) machining. This paper addresses the challenge of improving contour accuracy by proposing a novel contour error estimation and compensation method based on dynamic time warping (DTW). By incorporating time information and geometric characteristics of the machining path, the proposed method introduces a multiple-information fusion algorithm to define distance characteristics between the planned and actual trajectory sequences. This allows the calculation of a distortion path and the establishment of a mapping model between the two sequences. To mitigate the effect of DTW singularity on contour error estimation, a mapping model is established between line segments to determine reference points. The position contour error and the direction contour error of the five-axis tool are accurately estimated using segmented Hermite interpolation, and a spatial iterative learning framework is employed to compensate for them. Experimental results demonstrate the effectiveness of the proposed method in dealing with estimation errors in complex trajectories and its good performance in improving contour accuracy.Note to Practitioners—This paper proposes an effective strategy for the estimation of contour errors in five-axis machining. Currently, most methods for contour error estimation in five-axis machining are based on the nearest point principle. However, this approach fails to accurately estimate contour errors for complex trajectories with significant curvature variations, leading to ineffective contour error compensation in subsequent stages. Therefore, we introduce a contour error estimation algorithm based on DTW. This algorithm takes into account the time information and geometric features of the machining path. Experimental results validate the feasibility and advantages of this approach.

Recent years have witnessed the trend of enhancing recommender systems with large language models (LLMs), namely, LLMRec. A common way is to fine-tune the LLMs with the instruction data transformed from user behaviors, stimulating the recommendation ability of LLMs. Similar to traditional recommender systems, integrating user data into LLMs raises privacy concerns. Users desire a tool to erase the impacts of their sensitive data from the trained models. To meet this user demand, LLMRec unlearning becomes pivotal to enable the removal of unusable data (e.g.historical behaviors) from established LLMRec models. However, existing methods mostly focus on partition strategies and approximate unlearning. These methods are not well-suited for the unique characteristics of LLMRec due to computational costs or incomplete removal. In this study, we propose the Adapter Partition and Aggregation (APA) framework for exact and efficient LLMRec unlearning while maintaining recommendation performance. APA achieves this by retraining PEFT adapters using data partitioning, constructing adapters for partitioned training data shards, and retraining only the affected adapters. To preserve recommendation performance and avoid significant inference costs, APA incorporates balanced and heterogeneous data partitioning, and parameter-level adapter aggregation with sample-adaptive adapter attention for each testing sample. Extensive experiments demonstrate the effectiveness and efficiency of our method.

Recommendation unlearning is an emerging task to serve users for erasing unusable data (e.g., some historical behaviors) from a well-trained recommender model. Existing methods process unlearning requests by fully or partially retraining the model after removing the unusable data. However, these methods are impractical due to the high computation cost of full retraining and the highly possible performance damage of partial training. In this light, a desired recommendation unlearning method should obtain a similar model as full retraining in a more efficient manner, i.e., achieving complete, efficient and harmless unlearning. In this work, we propose a new Influence Function-based Recommendation Unlearning (IFRU) framework, which efficiently updates the model without retraining by estimating the influence of the unusable data on the model via the influence function . In the light that recent recommender models use historical data for both the constructions of the optimization loss and the computational graph (e.g., neighborhood aggregation), IFRU jointly estimates the direct influence of unusable data on optimization loss and the spillover influence on the computational graph to pursue complete unlearning. Furthermore, we propose an importance-based pruning algorithm to reduce the cost of the influence function. IFRU is harmless and applicable to mainstream differentiable models. Extensive experiments demonstrate that IFRU achieves more than 250 times acceleration compared to retraining-based methods with recommendation performance comparable to full retraining. Codes are available at https://github.com/baiyimeng/IFRU.