We introduce OSVBench, a new benchmark for evaluating Large Language Models (LLMs) on the task of generating complete formal specifications for verifying the functional correctness of operating system kernels. This benchmark is built upon a real-world operating system kernel, Hyperkernel, and consists of 245 complex specification generation tasks in total, each of which is a long-context task of about 20k-30k tokens. The benchmark formulates the specification generation task as a program synthesis problem confined to a domain for specifying states and transitions. This formulation is provided to LLMs through a programming model. The LLMs must be able to understand the programming model and verification assumptions before delineating the correct search space for syntax and semantics and generating formal specifications. Guided by the operating system's high-level functional description, the LLMs are asked to generate a specification that fully describes all correct states and transitions for a potentially buggy code implementation of the operating system. Experimental results with 12 state-of-the-art LLMs indicate limited performance of existing LLMs on the specification generation task for operating system verification. Significant disparities in their performance highlight differences in their ability to handle long-context code generation tasks.

Large Language Models (LLMs) have garnered remarkable advancements across diverse code-related tasks, known as Code LLMs, particularly in code generation that generates source code with LLM from natural language descriptions. This burgeoning field has captured significant interest from both academic researchers and industry professionals due to its practical significance in software development, e.g., GitHub Copilot . Despite the active exploration of LLMs for a variety of code tasks, either from the perspective of Natural Language Processing (NLP) or Software Engineering (SE) or both, there is a noticeable absence of a comprehensive and up-to-date literature review dedicated to LLM for code generation. In this survey, we aim to bridge this gap by providing a systematic literature review that serves as a valuable reference for researchers investigating the cutting-edge progress in LLMs for code generation. We introduce a taxonomy to categorize and discuss the recent developments in LLMs for code generation, covering aspects such as data curation, latest advances, performance evaluation, ethical implications, environmental impact, and real-world applications. In addition, we present a historical overview of the evolution of LLMs for code generation and provide a quantitative and qualitative comparative analysis of experimental results of code LLMs, sourced from their original papers to ensure a fair comparison on the HumanEval, MBPP, and BigCodeBench benchmarks, across various levels of difficulty and types of programming tasks, to highlight the progressive enhancements in LLM capabilities for code generation. We identify critical challenges and promising opportunities regarding the gap between academia and practical development. Furthermore, we have established a dedicated resource GitHub page ( https://github.com/juyongjiang/CodeLLMSurvey ) to continuously document and disseminate the most recent advances in the field.

The widespread adoption of cloud-based proprietary large language models (LLMs) has introduced significant challenges, including operational dependencies, privacy concerns, and the necessity of continuous internet connectivity.In this work, we introduce an LLMOps pipeline, "LlamaDuo", for the seamless migration of knowledge and abilities from serviceoriented LLMs to smaller, locally manageable models.This pipeline is crucial for ensuring service continuity in the presence of operational failures, strict privacy policies, or offline requirements.Our LlamaDuo involves fine-tuning a small language model against the service LLM using a synthetic dataset generated by the latter.If the performance of the fine-tuned model falls short of expectations, it is automatically improved through additional fine-tuning using extra similar data generated by the service LLM.This multi-turn process guarantees that the smaller model can eventually match or even surpass the service LLM's capabilities in specific downstream tasks, offering a practical and scalable solution for managing AI deployments in constrained environments.Extensive experiments with leadingedge LLMs are conducted to demonstrate the effectiveness, adaptability, and affordability of LlamaDuo across various downstream tasks.Our pipeline implementation is available at https://github.com/deep-diver/llamaduo.

The increasing sizes of large language models (LLMs) result in significant computational overhead and memory usage when adapting these models to specific tasks or domains. Various parameter-efficient fine-tuning (PEFT) methods have been devised to mitigate these challenges by training a small set of parameters for the task-specific updates of the model weights. Among PEFT methods, LoRA stands out for its simplicity and efficiency, inspiring the development of a series of variants. However, LoRA and its successors disregard the knowledge that is noisy or irrelevant to the targeted task, detrimentally impacting model performance and leading to suboptimality. To address this limitation, we introduce Knowledge-aware Singular-value Adaptation (KaSA), a PEFT method that leverages singular value decomposition (SVD) with knowledge-aware singular values to dynamically activate knowledge based on its relevance to the task at hand. We conduct extensive experiments across a range of LLMs on tasks spanning natural language understanding (NLU), generation (NLG), instruction following, and commonsense reasoning. The experimental results demonstrate that KaSA consistently outperforms FFT and 14 popular PEFT baselines across 16 benchmarks and 4 synthetic datasets, underscoring our method's efficacy and adaptability. The source code of our method is available at https://github.com/juyongjiang/KaSA.

Expert parallelism has emerged as a key strategy for distributing the computational workload of sparsely-gated mixture-of-experts (MoE) models across multiple devices, enabling the processing of increasingly large-scale models. However, the All-to-All communication inherent to expert parallelism poses a significant bottleneck, limiting the efficiency of MoE models. Although existing optimization methods partially mitigate this issue, they remain constrained by the sequential dependency between communication and computation operations. To address this challenge, we propose ScMoE, a novel shortcut-connected MoE architecture integrated with an overlapping parallelization strategy. ScMoE decouples communication from its conventional sequential ordering, enabling up to 100\% overlap with computation. Compared to the prevalent top-2 MoE baseline, ScMoE achieves speedups of $1.49\times$ in training and $1.82\times$ in inference. Moreover, our experiments and analyses indicate that ScMoE not only achieves comparable but in some instances surpasses the model quality of existing approaches.

Profile-guided optimization (PGO) advances the frontiers of compiler optimization by leveraging dynamic runtime information to generate highly optimized binaries. Traditional instrumentation-based profiling collects accurate profile data but often suffers from heavy runtime overhead. In contrast, sampling-based profiling is more efficient and scalable when collecting profile data while avoiding intrusive source code modifications. However, accurately collecting execution profiles via sampling remains challenging, especially when applied to fully optimized binaries. Such inaccurate profile data can restrict the benefits of PGO. This paper presents ProfiX, a machine learning-guided approach based on hybrid GNN architecture that addresses the problem of profile inference, aiming to correct inaccuracies in the profiles collected by sampling. Experiments on the SPEC 2017 benchmarks demonstrate that ProfiX achieves up to a 9.15\% performance improvement compared to the state-of-the-art traditional algorithm and an average 6.26\% improvement over the baseline machine learning models. These results highlight the effectiveness of ProfiX in optimizing real-world application profiles.

Large language models (LLMs) have garnered unprecedented advancements across diverse fields, ranging from natural language processing to computer vision and beyond. The prowess of LLMs is underpinned by their substantial model size, extensive and diverse datasets, and the vast computational power harnessed during training, all of which contribute to the emergent abilities of LLMs (e.g., in-context learning) that are not present in small models. Within this context, the mixture of experts (MoE) has emerged as an effective method for substantially scaling up model capacity with minimal computation overhead, gaining significant attention from academia and industry. Despite its growing prevalence, there lacks a systematic and comprehensive review of the literature on MoE. This survey seeks to bridge that gap, serving as an essential resource for researchers delving into the intricacies of MoE. We first briefly introduce the structure of the MoE layer, followed by proposing a new taxonomy of MoE. Next, we overview the core designs for various MoE models including both algorithmic and systemic aspects, alongside collections of available open-source implementations, hyperparameter configurations and empirical evaluations. Furthermore, we delineate the multifaceted applications of MoE in practice, and outline some potential directions for future research.

Sequential recommendation systems are integral to discerning temporal user preferences. Yet, the task of learning from abbreviated user interaction sequences poses a notable challenge. Data augmentation has been identified as a potent strategy to enhance the informational richness of these sequences. Traditional augmentation techniques, such as item randomization, may disrupt the inherent temporal dynamics. Although recent advancements in reverse chronological pseudo-item generation have shown promise, they can introduce temporal discrepancies when assessed in a natural chronological context. In response, we introduce a sophisticated approach, Bidirectional temporal data Augmentation with pre-training (BARec). Our approach leverages bidirectional temporal augmentation and knowledge-enhanced fine-tuning to synthesize authentic pseudo-prior items thatretain user preferences and capture deeper item semantic correlations, thus boosting the model’s expressive power. Our comprehensive experimental analysis on five benchmark datasets confirms the superiority of BARec across both short and elongated sequence contexts. Moreover, theoretical examination and case study offer further insight into the model’s logical processes and interpretability.

Convolutional neural networks (CNNs) have achieved significant performance on various real-life tasks. However, the large number of parameters in convolutional layers requires huge storage and computation resources, making it challenging to deploy CNNs on memory-constrained embedded devices. In this article, we propose a novel compression method that generates the convolution filters in each layer using a set of learnable low-dimensional quantized filter bases. The proposed method reconstructs the convolution filters by stacking the linear combinations of these filter bases. By using quantized values in weights, the compact filters can be represented using fewer bits so that the network can be highly compressed. Furthermore, we explore the sparsity of coefficients through $L_{1}$ -ball projection when conducting linear combination to further reduce the storage consumption and prevent overfitting. We also provide a detailed analysis of the compression performance of the proposed method. Evaluations of image classification and object detection tasks using various network structures demonstrate that the proposed method achieves a higher compression ratio with comparable accuracy compared with the existing state-of-the-art filter decomposition and network quantization methods.

As an effective tool for network compression, pruning techniques have been widely used to reduce the large number of parameters in deep neural networks (NNs). Nevertheless, unstructured pruning has the limitation of dealing with the sparse and irregular weights. By contrast, structured pruning can help eliminate this drawback but it requires complex criteria to determine which components to be pruned. Therefore, this paper presents a new method termed BUnit-Net, which directly constructs compact NNs by stacking designed basic units, without requiring additional judgement criteria anymore. Given the basic units of various architectures, they are combined and stacked systematically to build up compact NNs which involve fewer weight parameters due to the independence among the units. In this way, BUnit-Net can achieve the same compression effect as unstructured pruning while the weight tensors can still remain regular and dense. We formulate BUnit-Net in diverse popular backbones in comparison with the state-of-the-art pruning methods on different benchmark datasets. Moreover, two new metrics are proposed to evaluate the trade-off of compression performance. Experiment results show that BUnit-Net can achieve comparable classification accuracy while saving around 80% FLOPs and 73% parameters. That is, stacking basic units provides a new promising way for network compression.

Traffic forecasting is an essential problem in urban planning and computing. The complex dynamic spatial-temporal dependencies among traffic objects (e.g., sensors and road segments) have been calling for highly flexible models; unfortunately, sophisticated models may suffer from poor robustness especially in capturing the trend of the time series (1st-order derivatives with time), leading to unrealistic forecasts. To address the challenge of balancing dynamics and robustness, we propose TrendGCN, a new scheme that extends the flexibility of GCNs and the distribution-preserving capacity of generative and adversarial loss for handling sequential data with inherent statistical correlations. On the one hand, our model simultaneously incorporates spatial (node-wise) embeddings and temporal (time-wise) embeddings to account for heterogeneous space-and-time convolutions; on the other hand, it uses GAN structure to systematically evaluate statistical consistencies between the real and the predicted time series in terms of both the temporal trending and the complex spatial-temporal dependencies. Compared with traditional approaches that handle step-wise predictive errors independently, our approach can produce more realistic and robust forecasts. Experiments on six benchmark traffic forecasting datasets and theoretical analysis both demonstrate the superiority and the state-of-the-art performance of TrendGCN. Source code is available at https://github.com/juyongjiang/TrendGCN.

Sequential recommendation (SR) aims to model users' dynamic preferences from a series of interactions. A pivotal challenge in user modeling for SR lies in the inherent variability of user preferences. An effective SR model is expected to capture both the long-term and short-term preferences exhibited by users, wherein the former can offer a comprehensive understanding of stable interests that impact the latter. To more effectively capture such information, we incorporate locality inductive bias into the Transformer by amalgamating its global attention mechanism with a local convolutional filter, and adaptively ascertain the mixing importance on a personalized basis through layer-aware adaptive mixture units, termed as AdaMCT. Moreover, as users may repeatedly browse potential purchases, it is expected to consider multiple relevant items concurrently in long-/short-term preferences modeling. Given that softmax-based attention may promote unimodal activation, we propose the Squeeze-Excitation Attention (with sigmoid activation) into SR models to capture multiple pertinent items (keys) simultaneously. Extensive experiments on three widely employed benchmarks substantiate the effectiveness and efficiency of our proposed approach. Source code is available at https://github.com/juyongjiang/AdaMCT.

Deep neural networks frequently suffer from performance degradation when the training data is long-tailed because several majority classes dominate the training, resulting in a biased model. Recent studies have made a great effort in solving this issue by obtaining good representations from data space, but few of them pay attention to the influence of feature norm on the predicted results. In this paper, we therefore address the long-tailed problem from feature space and thereby propose the feature-balanced loss. Specifically, we encourage larger feature norms of tail classes by giving them relatively stronger stimuli. Moreover, the stimuli intensity is gradually increased in the way of curriculum learning, which improves the generalization of the tail classes, meanwhile maintaining the performance of the head classes. Extensive experiments on multiple popular long-tailed recognition benchmarks demonstrate that the feature-balanced loss achieves superior performance gains compared with the state-of-the-art methods.