Model merging has emerged as a promising technique for enhancing large language models, though its application in large-scale pre-training remains relatively unexplored. In this paper, we present a comprehensive investigation of model merging techniques during the pre-training process. Through extensive experiments with both dense and Mixture-of-Experts (MoE) architectures ranging from millions to over 100 billion parameters, we demonstrate that merging checkpoints trained with constant learning rates not only achieves significant performance improvements but also enables accurate prediction of annealing behavior. These improvements lead to both more efficient model development and significantly lower training costs. Our detailed ablation studies on merging strategies and hyperparameters provide new insights into the underlying mechanisms while uncovering novel applications. Through comprehensive experimental analysis, we offer the open-source community practical pre-training guidelines for effective model merging.

Large Language Models (LLMs) have been suggested for use in automated vulnerability repair, but benchmarks showing they can consistently identify security-related bugs are lacking. We thus develop SecLLMHolmes, a fully automated evaluation framework that performs the most detailed investigation to date on whether LLMs can reliably identify and reason about security-related bugs. We construct a set of 228 code scenarios and analyze eight of the most capable LLMs across eight different investigative dimensions using our framework. Our evaluation shows LLMs provide non-deterministic responses, incorrect and unfaithful reasoning, and perform poorly in real-world scenarios. Most importantly, our findings reveal significant non-robustness in even the most advanced models like ‘PaLM2’ and ‘GPT-4’: by merely changing function or variable names, or by the addition of library functions in the source code, these models can yield incorrect answers in 26% and 17% of cases, respectively. These findings demonstrate that further LLM advances are needed before LLMs can be used as general purpose security assistants.

A knowledge graph contains many real-world facts that can be used to support various analytical tasks, e.g., exceptional fact discovery and the check of claims. In this work, we attempt to extract top-$k$frequent and diversified patterns from knowledge graph by well capturing user interest. Specifically, we first formalize the core-based top-$k$frequent pattern discovery problem, which finds the top-$k$frequent patterns that are extended from a core pattern specified by user query and have the highest frequency. In addition, to diversify the top-$k$frequent patterns, we define a distance function to measure the dissimilarity between two patterns, and return top-$k$patterns in which the pairwise diversity of any two resultant patterns exceeds a given threshold. As the search space of candidate patterns is exponential w.r.t. the number of nodes and edges in the knowledge graph, discovering frequent and diversified patterns is computationally challenging. To achieve high efficiency, we propose a suite of techniques, including (1) We devise a meta-index to avoid generating invalid candidate patterns; (2) We propose an upper bound of the frequency score (i.e., MNI) of the candidate pattern, which is used to prune unqualified candidates earlier and prioritize the enumeration order of patterns; (3) We design an advanced join-based approach to compute the MNI of candidate patterns efficiently; and (4) We develop a lower bound for distance function and incrementally compute the pairwise diversity among the patterns. Using real-world knowledge graphs, we experimentally verify the efficiency and effectiveness of our proposed techniques. We also demonstrate the utility of the extracted patterns by case studies.

As a popular distributed data warehouse system, Apache Hive has been widely used for big data analytics in many organizations. Meanwhile, exploiting the massive parallelism of GPU to accelerate online analytical processing (OLAP) has been extensively explored in the database community. In this paper, we present GHive, which enhances CPU-based Hive via CPU-GPU heterogeneous computing. GHive is designed for the business intelligence applications and provides the same API as Hive for compatibility. To run SQL queries jointly on both CPU and GPU, GHive comes with three key techniques: (i) a novel data model gTable, which is column-based and enables efficient data movement between CPU memory and GPU memory; (ii) a GPU-based operator library Panda, which provides a complete set of SQL operators with extensively optimized GPU implementations; (iii) a hardware-aware MapReduce job placement scheme, which puts jobs judiciously on either GPU or CPU via a cost-based approach. In the experiments, we observe that GHive outperforms Hive in both query processing speed and operating expense on the Star Schema Benchmark (SSB).

As a distributed, fault-tolerant data warehouse system for large-scale data analytics, Apache Hive has been used for various applications in many organizations (e.g., Facebook, Amazon, and Huawei). Exploiting the large degrees of parallelism of GPU to improve the performance of online analytical processing (OLAP) in database system is a common practice in the industry. Meanwhile, it is a common practice to exploit the large degrees of parallelism of GPU to improve the performance of online analytical processing (OLAP) in database systems. This demo presents GHive, which enables Apache Hive to accelerate OLAP queries by jointly utilizing CPU and GPU in intelligent and efficient ways. The takeaways for SIGMOD attendees include: (1) the superior performance of GHive compared with vanilla Hive that only uses CPU; (2) intuitive visualizations of execution statistics for Hive and GHive to understand where the acceleration of GHive comes from; (3) detailed profiling of the time taken by each operator on CPU and GPU to show the advantages of GPU execution.

Knowledge graph is a way of structuring information in graph form, by representing entities as nodes and relationships between entities as edges. A knowledge graph often consists of large amount of facts in real-world which can be used in supporting many analytical tasks, e.g., exceptional facts discovery and fact check of claims. In this work, we study a core-based top-k frequent pattern discovery problem which is frequently used as a subroutine in analyzing knowledge graphs. The main challenge of the problem is search space of the candidate patterns is exponential to the combinations of the nodes and edges in the knowledge graph.To reduce the search space, we devise a novel computation framework FastPat with a suite of optimizations. First, we devise a meta-index, which can be used to avoid generating invalid candidate patterns. Second, we propose an upper bound of the frequency score (i.e., MNI) of the candidate pattern that prunes unqualified candidates earlier and prioritize the enumeration order of the patterns. Lastly, we design a join-based approach to compute the MNI of candidate pattern efficiently. We conduct extensive experimental studies in real-world datasets to verify the superiority of our proposed method over the baselines. We also demonstrate the utility of the discovered frequent patterns by a case study in COVID-19 knowledge graph.

In this paper, we study the m-impact region problem (mIR). In a context where users look for available products with top-k queries, mIR identifies the part of the product space that attracts the most user attention. Specifically, mIR determines the kind of attribute values that lead a (new or existing) product to the top-k result for at least a fraction of the user population. mIR has several applications, ranging from effective marketing to product improvement. Importantly, it also leads to (exact and efficient) solutions for standing top-k impact problems, which were previously solved heuristically only, or whose current solutions face serious scalability limitations. We experiment, among others, on data mined from actual user reviews for real products, and demonstrate the practicality and efficiency of our algorithms, both for mIR and for standing top-k impact problems.