Conditional molecular generation, aiming to generate 2D and 3D molecules that satisfy given properties, has achieved remarkable progress, thanks to the advances in deep generative models such as graph diffusion. However, existing methods generally assume that the given conditions for training and testing are consistent, failing to handle the realistic challenge when there exist distribution shifts between training and testing conditions. Invariant learning is a mainstream paradigm for addressing distribution shifts, but fusing invariant learning principles with conditional molecular generation faces three core challenges: (1) existing invariant learning methods focus on discriminative tasks and cannot be directly adapted to molecule generative tasks; (2) how to distinguish between invariant subgraph and variant subgraph of a molecule graph, which is treated as an integrated input; (3) how to fuse invariant subgraphs, variant subgraphs, and property conditions for effective generation. To tackle these challenges, we propose Invariant Conditional MOLecular generation (IC-MOL), a framework that combines invariant learning with graph diffusion to improve the generalization ability of conditional molecular generation under distribution shifts. Specifically, we first disentangle molecular graphs into invariant and variant subgraphs while maintaining SE(3) equivariance, an important inductive bias for molecular generation. On this basis, we further design a two-phase graph diffusion generation model. In the first phase, we generate an invariant molecular consistent with the target property. In the second phase, we propose a cross-attention mechanism to fuse variant subgraph representations and property conditions to guide the generation of complete molecules while maintaining property alignment. Extensive experiments on the benchmark dataset show that IC-MOL consistently outperforms state-of-the-art baselines across six property conditions under distribution shifts.

CUDA kernels are essential for high-performance computing, yet their deployment has been limited to vendor-specific chips. Unless highly efficient computational kernels are custom-implemented by experts, other chips may face problems such as low utilization and inability to accelerate AI computing. In this paper, we introduce CKTI, a domain-specific compiler that lowers CUDA kernels to Triton-IR, thereby decoupling them from proprietary hardware and fostering diversity across the computing ecosystem. CKTI proposes a scheduling algorithm that transforms the threading model from thread-level to tile-level, along with a mapping scheme from explicit control-structure characteristics to dynamic masks. It also incorporates a custom dialect to express the complete semantics and optimizable properties of the kernel. These collectively ensure performance portability and cross-platform deployment. The results show that CKTI produces correct outputs on different hardware platforms and delivers competitive performance, achieving 1.28X speedup on NVIDIA, 1.17X on AMD, 1.14X on MetaX, and unlocking deployment on Cambricon platforms. Additionally, we validate CKTI’s support for end-to-end workloads across multiple architecture, including NPUs and GPGPUs.

For large IT companies, maintaining numerous software systems presents considerable complexity. Logs are invaluable for depicting the state of systems, making log-based anomaly detection crucial for ensuring system reliability. Existing methods require extensive log data for training, hindering their rapid deployment for new systems. Cross-system log anomaly detection methods attempt to transfer knowledge from mature systems to new ones but often struggle with syntax differences and system-specific knowledge, which hinders their effectiveness. To address these issues, this paper proposes LogSynergy, a novel transfer learning-based log anomaly detection framework. LogSynergy employs (1) LLM-based event interpretation (LEI) to standardize log syntax across different systems, and (2) system-unified feature extraction (SUFE) to disentangle system-specific features from system-unified features. These bridge the gap among different systems and enhance LogSynergy's generalizability. LogSynergy has been deployed in the production environment of a top-tier global Internet Service Provider (ISP), where it was evaluated on three real-world datasets. Additionally, we conducted evaluations on three public datasets. The results demonstrate that LogSynergy significantly outperforms existing methods. It achieves F1-scores over 89% on the real-world datasets and over 83% on the public datasets, using only 5000 labeled log sequences from the new system. These results underscore LogSynergy's effectiveness in rapidly deploying anomaly detection models for new systems. The code of LogSynergy has been open-sourced at https://github.com/DDUtian/LogSynergy

Deep learning compilers optimize DNN program execution by capturing them as operator-based computation graphs. However, developers’ deep learning programs often contain complex Python language features that prevent compilers from recognizing the entire program as a complete computation graph, resulting in sub-optimal performance. Our analysis reveals that actual capture failures involve only a few lines of code, we believe this problem can be addressed through code repair rather than extensive compiler improvements. To address this challenge, we introduce GraphGlue, a multi-agent system that leverages LLMs to repair and optimize DNN programs for compiler requirements, thereby maximizing the performance benefits of deep learning compilers in inference scenarios. GraphGlue employs (1) graph-break cause mining (GCM) to identify hidden causes of computation graph breaks and facilitate LLM-based repair, and (2) self-correction with reject sampling (SRS) to alternate between code debugging and regeneration, effectively avoiding ineffective feedback attempts caused by incorrect initial optimization strategies. Experimental results demonstrate that programs optimized by GraphGlue achieve up to 2.19x (1.23x on average) speedup compared to using TorchDynamo directly, and deliver up to 15.77x (8.74x on average) memory savings compared to state-of-the-art AI compiler frontends. GraphGlue exhibits strong generalization capabilities across 1,411 real-world user programs, successfully optimizing 92.63% of them. Code is available at https://github.com/Jamesswang/GraphGlue.

With the rising demand for computational power and the increasing variety of computational scenarios, considerable interest has emerged in transforming existing CUDA programs into more general-purpose OpenCL programs, enabling them to run across diverse hardware platforms. However, manual methods, typically designed for specific applications, lack flexibility. Current automated conversion techniques also face considerable challenges, particularly in handling diverse programming interfaces, memory management, and so on, and are insufficient for converting large-scale, complex CUDA projects. In this article, we propose a novel source-to-source program transformation framework, TransCL, which automates the conversion of CUDA programs in four key aspects: source code, execution model, programming model, and memory model. To achieve this, we abstract a set of conversion rules aligned with the latest CUDA standards, develop a transcoder, implement an OpenCL-compatible programming interface library, and establish a memory mapping mechanism between CUDA and OpenCL. Experiments demonstrate that TransCL provides a high level of automation in converting CUDA-based applications and is effective in handling large, complex projects such as TensorFlow. Moreover, the converted AI framework successfully conducted model training for the first time. The experiment also validates that the converted program can execute correctly across multiple platforms and demonstrate good performance.

Logs are one of the most valuable data to describe the running state of services. Failure diagnosis through logs is crucial for service reliability and security. The current automatic log failure diagnosis methods cannot fully use the multiple fields of logs, which fail to capture the relation between them. In this article, we propose LogKG, a new framework for diagnosing failures based on knowledge graphs (KG) of logs. LogKG fully extracts entities and relations from logs to mine multi-field information and their relations through the KG. To fully use the information represented by KG, we propose a failure-oriented log representation (FOLR) method to extract the failure-related patterns. Utilizing the OPTICS clustering method, LogKG aggregates historical failure cases, labels typical failure cases, and trains a failure diagnosis model to identify the root cause. We evaluate the effectiveness of LogKG on a real-world log dataset and a public log dataset, respectively, showing that it outperforms existing methods. With the deployment in a top-tier global Internet Service Provider (ISP), we demonstrate the performance and practicability of LogKG.

To ensure the reliability of software services, operators collect and monitor a large number of KPI (Key Performance Indicator) streams constantly. KPI anomaly detection is vitally important for software service management. However, none of supervised learning methods, semi-supervised learning methods, transfer learning methods, or unsupervised learning methods achieve accurate anomaly detection for the large-scale, diverse, dynamically changing KPI streams with little labeling effort. In this paper, we propose PUAD, a PU learning-based method, to achieve accurate KPI anomaly detection requiring a few partial labels. It integrates clustering, PU learning, and semi-supervised learning to minimize labeling effort and improve anomaly detection accuracy simultaneously. Additionally, we propose a novel active learning method that selects the samples most likely to be positive in each iteration to avoid false alarms. We apply 208 real-world KPI streams collected from a large-scale software service provider to evaluate the performance of PUAD, demonstrating that it achieves a close F1-score to supervised learning methods with much fewer manual labels, and greatly outperforms semi-supervised learning methods, transfer learning methods, and unsupervised learning methods.