This paper systematically explores how Large Language Models (LLMs) generate explanations of code examples of the type used in intro-to-programming courses. As we show, the nature of code explanations generated by LLMs varies considerably based on the wording of the prompt, the target code examples being explained, the programming language, the temperature parameter, and the version of the LLM. Nevertheless, they are consistent in two major respects for Java and Python: the readability level, which hovers around 7-8 grade, and lexical density, i.e., the relative size of the meaningful words with respect to the total explanation size. Furthermore, the explanations score very high in correctness but less on three other metrics: completeness, conciseness, and contextualization.
相關內容
This paper studies the Partial Optimal Transport (POT) problem between two unbalanced measures with at most $n$ supports and its applications in various AI tasks such as color transfer or domain adaptation. There is hence the need for fast approximations of POT with increasingly large problem sizes in arising applications. We first theoretically and experimentally investigate the infeasibility of the state-of-the-art Sinkhorn algorithm for POT due to its incompatible rounding procedure, which consequently degrades its qualitative performance in real world applications like point-cloud registration. To this end, we propose a novel rounding algorithm for POT, and then provide a feasible Sinkhorn procedure with a revised computation complexity of $\mathcal{\widetilde O}(n^2/\varepsilon^4)$. Our rounding algorithm also permits the development of two first-order methods to approximate the POT problem. The first algorithm, Adaptive Primal-Dual Accelerated Gradient Descent (APDAGD), finds an $\varepsilon$-approximate solution to the POT problem in $\mathcal{\widetilde O}(n^{2.5}/\varepsilon)$, which is better in $\varepsilon$ than revised Sinkhorn. The second method, Dual Extrapolation, achieves the computation complexity of $\mathcal{\widetilde O}(n^2/\varepsilon)$, thereby being the best in the literature. We further demonstrate the flexibility of POT compared to standard OT as well as the practicality of our algorithms on real applications where two marginal distributions are unbalanced.
This paper explores the intricacies of traffic behavior at unsignalized intersections through the lens of a novel dataset, combining manual video data labeling and advanced traffic simulation in SUMO. This research involved recording traffic at various unsignalized intersections in Memphis, TN, during different times of the day. After manually labeling video data to capture specific variables, we reconstructed traffic scenarios in the SUMO simulation environment. The output data from these simulations offered a comprehensive analysis, including time-space diagrams for vehicle movement, travel time frequency distributions, and speed-position plots to identify bottleneck points. This approach enhances our understanding of traffic dynamics, providing crucial insights for effective traffic management and infrastructure improvements.
This paper establishes the equivalence between Local Differential Privacy (LDP) and a global limit on learning any knowledge about an object. However, an output from an LDP query is not necessarily required to provide exact amount of knowledge equal to the upper bound of the learning limit. Since the amount of knowledge gain should be proportional to the incurred privacy loss, the traditional approach of using DP guarantee to measure privacy loss can occasionally overestimate the actual privacy loss. This is especially problematic in privacy accounting in LDP, where privacy loss is computed by accumulating the DP guarantees. To address this issue, this paper introduces the concept of \textit{realized privacy loss}, which measures the actual knowledge gained by the analyst after a query, as a more accurate measure of privacy loss. The realized privacy loss is integrated into the privacy accounting of fully adaptive composition, where an adversary adaptively selects queries based on previous results. Bayesian Privacy Filter is implemented to continually accept queries until the realized privacy loss of the composed queries equals the DP guarantee of the composition, allowing the full utilization of the privacy budget. Tracking the realized privacy loss during the composition is achieved through Bayesian Privacy Odometer, and the gap between the privacy budget and the realized privacy loss measures the leeway of the DP guarantee for future queries. A branch-and-bound method is devised to enable the Bayesian Privacy Filter to safeguard objects with continuous values. The Bayesian Privacy Filter is proven to be at least as efficient as the basic composition, and more efficient if the queries are privacy-loss compactible. Experimental results indicate that Bayesian Privacy Filter outperforms the basic composition by a factor of one to four when composing linear and logistic regressions.
This paper studies safe Reinforcement Learning (safe RL) with linear function approximation and under hard instantaneous constraints where unsafe actions must be avoided at each step. Existing studies have considered safe RL with hard instantaneous constraints, but their approaches rely on several key assumptions: $(i)$ the RL agent knows a safe action set for {\it every} state or knows a {\it safe graph} in which all the state-action-state triples are safe, and $(ii)$ the constraint/cost functions are {\it linear}. In this paper, we consider safe RL with instantaneous hard constraints without assumption $(i)$ and generalize $(ii)$ to Reproducing Kernel Hilbert Space (RKHS). Our proposed algorithm, LSVI-AE, achieves $\tilde{\cO}(\sqrt{d^3H^4K})$ regret and $\tilde{\cO}(H \sqrt{dK})$ hard constraint violation when the cost function is linear and $\cO(H\gamma_K \sqrt{K})$ hard constraint violation when the cost function belongs to RKHS. Here $K$ is the learning horizon, $H$ is the length of each episode, and $\gamma_K$ is the information gain w.r.t the kernel used to approximate cost functions. Our results achieve the optimal dependency on the learning horizon $K$, matching the lower bound we provide in this paper and demonstrating the efficiency of LSVI-AE. Notably, the design of our approach encourages aggressive policy exploration, providing a unique perspective on safe RL with general cost functions and no prior knowledge of safe actions, which may be of independent interest.
This paper concerns the development of metatheory for extensible languages. It uses as its starting point a view that programming languages tailored to specific application domains are to be constructed by composing components from an open library of independently-developed extensions to a host language. In the elaboration of this perspective, static analyses (such as typing) and dynamic semantics (such as evaluation) are described via relations whose specifications are distributed across the host language and extensions and are given in a rule-based fashion. Metatheoretic properties, which ensure that static analyses accurately gauge runtime behavior, are represented in this context by formulas over such relations. These properties may be fundamental to the language, introduced by the host language, or they may pertain to analyses introduced by individual extensions. We expose the problem of modular metatheory, i.e., the notion that proofs of relevant properties can be constructed by reasoning independently within each component in the library. To solve this problem, we propose the twin ideas of decomposing proofs around language fragments and of reasoning generically about extensions based on broad, a priori constraints imposed on their behavior. We establish the soundness of these styles of reasoning by showing how complete proofs of the properties can be automatically constructed for any language obtained by composing the independent parts. Mathematical precision is given to our discussions by framing them within a logic that encodes inductive rule-based specifications via least fixed-point definitions. We also sketch the structure of a practical system for metatheoretic reasoning for extensible languages based on the ideas developed.
This paper presents a detailed case study examining the application of Large Language Models (LLMs) in the construction of test cases within the context of software engineering. LLMs, characterized by their advanced natural language processing capabilities, are increasingly garnering attention as tools to automate and enhance various aspects of the software development life cycle. Leveraging a case study methodology, we systematically explore the integration of LLMs in the test case construction process, aiming to shed light on their practical efficacy, challenges encountered, and implications for software quality assurance. The study encompasses the selection of a representative software application, the formulation of test case construction methodologies employing LLMs, and the subsequent evaluation of outcomes. Through a blend of qualitative and quantitative analyses, this study assesses the impact of LLMs on test case comprehensiveness, accuracy, and efficiency. Additionally, delves into challenges such as model interpretability and adaptation to diverse software contexts. The findings from this case study contributes with nuanced insights into the practical utility of LLMs in the domain of test case construction, elucidating their potential benefits and limitations. By addressing real-world scenarios and complexities, this research aims to inform software practitioners and researchers alike about the tangible implications of incorporating LLMs into the software testing landscape, fostering a more comprehensive understanding of their role in optimizing the software development process.
Denoising Diffusion models have exhibited remarkable capabilities in image generation. However, generating high-quality samples requires a large number of iterations. Knowledge distillation for diffusion models is an effective method to address this limitation with a shortened sampling process but causes degraded generative quality. Based on our analysis with bias-variance decomposition and experimental observations, we attribute the degradation to the spatial fitting error occurring in the training of both the teacher and student model. Accordingly, we propose $\textbf{S}$patial $\textbf{F}$itting-$\textbf{E}$rror $\textbf{R}$eduction $\textbf{D}$istillation model ($\textbf{SFERD}$). SFERD utilizes attention guidance from the teacher model and a designed semantic gradient predictor to reduce the student's fitting error. Empirically, our proposed model facilitates high-quality sample generation in a few function evaluations. We achieve an FID of 5.31 on CIFAR-10 and 9.39 on ImageNet 64$\times$64 with only one step, outperforming existing diffusion methods. Our study provides a new perspective on diffusion distillation by highlighting the intrinsic denoising ability of models. Project link: \url{//github.com/Sainzerjj/SFERD}.
This paper offers a comprehensive review of the research on Natural Language Generation (NLG) over the past two decades, especially in relation to data-to-text generation and text-to-text generation deep learning methods, as well as new applications of NLG technology. This survey aims to (a) give the latest synthesis of deep learning research on the NLG core tasks, as well as the architectures adopted in the field; (b) detail meticulously and comprehensively various NLG tasks and datasets, and draw attention to the challenges in NLG evaluation, focusing on different evaluation methods and their relationships; (c) highlight some future emphasis and relatively recent research issues that arise due to the increasing synergy between NLG and other artificial intelligence areas, such as computer vision, text and computational creativity.
In this paper, we propose a novel Feature Decomposition and Reconstruction Learning (FDRL) method for effective facial expression recognition. We view the expression information as the combination of the shared information (expression similarities) across different expressions and the unique information (expression-specific variations) for each expression. More specifically, FDRL mainly consists of two crucial networks: a Feature Decomposition Network (FDN) and a Feature Reconstruction Network (FRN). In particular, FDN first decomposes the basic features extracted from a backbone network into a set of facial action-aware latent features to model expression similarities. Then, FRN captures the intra-feature and inter-feature relationships for latent features to characterize expression-specific variations, and reconstructs the expression feature. To this end, two modules including an intra-feature relation modeling module and an inter-feature relation modeling module are developed in FRN. Experimental results on both the in-the-lab databases (including CK+, MMI, and Oulu-CASIA) and the in-the-wild databases (including RAF-DB and SFEW) show that the proposed FDRL method consistently achieves higher recognition accuracy than several state-of-the-art methods. This clearly highlights the benefit of feature decomposition and reconstruction for classifying expressions.
In this paper, we proposed to apply meta learning approach for low-resource automatic speech recognition (ASR). We formulated ASR for different languages as different tasks, and meta-learned the initialization parameters from many pretraining languages to achieve fast adaptation on unseen target language, via recently proposed model-agnostic meta learning algorithm (MAML). We evaluated the proposed approach using six languages as pretraining tasks and four languages as target tasks. Preliminary results showed that the proposed method, MetaASR, significantly outperforms the state-of-the-art multitask pretraining approach on all target languages with different combinations of pretraining languages. In addition, since MAML's model-agnostic property, this paper also opens new research direction of applying meta learning to more speech-related applications.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191