In Large Language Models (LLMs), there have been consistent advancements in task-specific performance, largely influenced by effective prompt design. While recent research on prompting has enhanced the reasoning capabilities of LLMs, a gap remains in further improving their understanding abilities. In this study, we introduce metacognitive prompting (MP), a strategy inspired by human introspective reasoning processes. Using MP, LLMs undergo a systematic series of structured, self-aware evaluations, drawing on both their vast inherent knowledge and new insights. Our experiments involve five prevalent LLMs: Llama2, Vicuna, PaLM, GPT-3.5, and GPT-4, all of which span various general natural language understanding (NLU) tasks from the GLUE and SuperGLUE benchmarks. Results indicate that, although GPT-4 consistently excels in most tasks, PaLM, when equipped with MP, approaches its performance level. Furthermore, across models and datasets, MP consistently outperforms existing prompting methods, including standard and chain-of-thought prompting. This study underscores the potential to amplify the understanding abilities of LLMs and highlights the benefits of mirroring human introspective reasoning in NLU tasks.
相關內容
Large Language Models (LLMs) have shown promise in multiple software engineering tasks including code generation, program repair, code summarisation, and test generation. Fault localisation is instrumental in enabling automated debugging and repair of programs and was prominently featured as a highlight during the launch event of ChatGPT-4. Nevertheless, the performance of LLMs compared to state-of-the-art methods, as well as the impact of prompt design and context length on their efficacy, remains unclear. To fill this gap, this paper presents an in-depth investigation into the capability of ChatGPT-3.5 and ChatGPT-4, the two state-of-the-art LLMs, on fault localisation. Using the widely-adopted large-scale Defects4J dataset, we compare the two LLMs with the existing fault localisation techniques. We also investigate the consistency of LLMs in fault localisation, as well as how prompt engineering and the length of code context affect the fault localisation effectiveness. Our findings demonstrate that within function-level context, ChatGPT-4 outperforms all the existing fault localisation methods. Additional error logs can further improve ChatGPT models' localisation accuracy and consistency, with an average 46.9% higher accuracy over the state-of-the-art baseline SmartFL on the Defects4J dataset in terms of TOP-1 metric. However, when the code context of the Defects4J dataset expands to the class-level, ChatGPT-4's performance suffers a significant drop, with 49.9% lower accuracy than SmartFL under TOP-1 metric. These observations indicate that although ChatGPT can effectively localise faults under specific conditions, limitations are evident. Further research is needed to fully harness the potential of LLMs like ChatGPT for practical fault localisation applications.
Graph Neural Networks (GNNs) have achieved promising performance in a variety of graph-focused tasks. Despite their success, existing GNNs suffer from two significant limitations: a lack of interpretability in results due to their black-box nature, and an inability to learn representations of varying orders. To tackle these issues, we propose a novel Model-agnostic Graph Neural Network (MaGNet) framework, which is able to sequentially integrate information of various orders, extract knowledge from high-order neighbors, and provide meaningful and interpretable results by identifying influential compact graph structures. In particular, MaGNet consists of two components: an estimation model for the latent representation of complex relationships under graph topology, and an interpretation model that identifies influential nodes, edges, and important node features. Theoretically, we establish the generalization error bound for MaGNet via empirical Rademacher complexity, and showcase its power to represent layer-wise neighborhood mixing. We conduct comprehensive numerical studies using simulated data to demonstrate the superior performance of MaGNet in comparison to several state-of-the-art alternatives. Furthermore, we apply MaGNet to a real-world case study aimed at extracting task-critical information from brain activity data, thereby highlighting its effectiveness in advancing scientific research.
Given the success of Large Language Models (LLMs), there has been considerable interest in studying the properties of model activations. The literature overwhelmingly agrees that LLM representations are dominated by a few ``outlier dimensions'' with exceedingly high variance and magnitude. Several studies in Natural Language Processing (NLP) have sought to mitigate the impact of such outlier dimensions and force LLMs to be isotropic (i.e., have uniform variance across all dimensions in embedding space). Isotropy is thought to be a desirable property for LLMs that improves model performance and more closely aligns textual representations with human intuition. However, many of the claims regarding isotropy in NLP have been based on the average cosine similarity of embeddings, which has recently been shown to be a flawed measure of isotropy. In this paper, we propose I-STAR: IsoScore*-based STable Anisotropic Regularization, a novel regularization method that can be used to increase or decrease levels of isotropy in embedding space during training. I-STAR uses IsoScore*, the first accurate measure of isotropy that is both differentiable and stable on mini-batch computations. In contrast to several previous works, we find that decreasing isotropy in contextualized embeddings improves performance on the majority of tasks and models considered in this paper.
Trusted Platform Modules (TPMs), which serve as the root of trust in secure systems, are secure crypto-processors that carry out cryptographic primitives. Should large-scale quantum computing become a reality, the cryptographic primitives adopted in the TPM 2.0 standard will no longer be secure. Thus, the design of TPMs that provide Quantum Resistant (QR) primitives is of utmost importance, in particular with the restrictions imposed by embedded systems. In this paper, we investigate the deployment of QR primitives and protocols in the standard TPM 2.0. Cryptographic algorithms that are already in the NIST QR cryptography standardization process, as well as an Oblivious Transfer (OT), a fundamental cryptographic primitive, are the QR cryptographic schemes selected to extend TPM 2.0. In particular, the Kyber algorithm for key encapsulation, the Dilithium algorithm for digital signature, and a 3-round Random Oblivious Transfer (ROT) protocol, supporting protocols such as Multi-Party Computation and Private Set Intersection (PSI). The QR extended TPM 2.0 is implemented in ARM and RISC-V embedded processors, its computational requirements are analysed and experimentally evaluated in comparison to the standard TPM. It is shown that Kyber and Dilithium are faster at creating keys than RSA, due to the key size and secure random sampling required in RSA, while they meet the same performance level as ECC. For digital signatures, both in signature creation and verification, Dilithium is on par with RSA and ECC. The ROT protocol shows decent performance and its support required small modifications to the TPM. This paper also shows that it would be possible to backport the required code to already available TPMs to ensure that current TPMs remain secure against quantum adversaries.
Large Language Models (LLMs) are quickly becoming ubiquitous, but the implications for social science research are not yet well understood. This paper asks whether LLMs can help us analyse large-N qualitative data from open-ended interviews, with an application to transcripts of interviews with Rohingya refugees in Cox's Bazaar, Bangladesh. We find that a great deal of caution is needed in using LLMs to annotate text as there is a risk of introducing biases that can lead to misleading inferences. We here mean bias in the technical sense, that the errors that LLMs make in annotating interview transcripts are not random with respect to the characteristics of the interview subjects. Training simpler supervised models on high-quality human annotations with flexible coding leads to less measurement error and bias than LLM annotations. Therefore, given that some high quality annotations are necessary in order to asses whether an LLM introduces bias, we argue that it is probably preferable to train a bespoke model on these annotations than it is to use an LLM for annotation.
Large Language Models (LLMs) have recently been shown to be effective as automatic evaluators with simple prompting and in-context learning. In this work, we assemble 15 LLMs of four different size ranges and evaluate their output responses by preference ranking from the other LLMs as evaluators, such as System Star is better than System Square. We then evaluate the quality of ranking outputs introducing the Cognitive Bias Benchmark for LLMs as Evaluators (CoBBLEr), a benchmark to measure six different cognitive biases in LLM evaluation outputs, such as the Egocentric bias where a model prefers to rank its own outputs highly in evaluation. We find that LLMs are biased text quality evaluators, exhibiting strong indications on our bias benchmark (average of 40% of comparisons across all models) within each of their evaluations that question their robustness as evaluators. Furthermore, we examine the correlation between human and machine preferences and calculate the average Rank-Biased Overlap (RBO) score to be 49.6%, indicating that machine preferences are misaligned with humans. According to our findings, LLMs may still be unable to be utilized for automatic annotation aligned with human preferences. Our project page is at: //minnesotanlp.github.io/cobbler.
While Reinforcement Learning (RL) achieves tremendous success in sequential decision-making problems of many domains, it still faces key challenges of data inefficiency and the lack of interpretability. Interestingly, many researchers have leveraged insights from the causality literature recently, bringing forth flourishing works to unify the merits of causality and address well the challenges from RL. As such, it is of great necessity and significance to collate these Causal Reinforcement Learning (CRL) works, offer a review of CRL methods, and investigate the potential functionality from causality toward RL. In particular, we divide existing CRL approaches into two categories according to whether their causality-based information is given in advance or not. We further analyze each category in terms of the formalization of different models, ranging from the Markov Decision Process (MDP), Partially Observed Markov Decision Process (POMDP), Multi-Arm Bandits (MAB), and Dynamic Treatment Regime (DTR). Moreover, we summarize the evaluation matrices and open sources while we discuss emerging applications, along with promising prospects for the future development of CRL.
In the era of deep learning, modeling for most NLP tasks has converged to several mainstream paradigms. For example, we usually adopt the sequence labeling paradigm to solve a bundle of tasks such as POS-tagging, NER, Chunking, and adopt the classification paradigm to solve tasks like sentiment analysis. With the rapid progress of pre-trained language models, recent years have observed a rising trend of Paradigm Shift, which is solving one NLP task by reformulating it as another one. Paradigm shift has achieved great success on many tasks, becoming a promising way to improve model performance. Moreover, some of these paradigms have shown great potential to unify a large number of NLP tasks, making it possible to build a single model to handle diverse tasks. In this paper, we review such phenomenon of paradigm shifts in recent years, highlighting several paradigms that have the potential to solve different NLP tasks.
Graph Neural Networks (GNNs) have recently been used for node and graph classification tasks with great success, but GNNs model dependencies among the attributes of nearby neighboring nodes rather than dependencies among observed node labels. In this work, we consider the task of inductive node classification using GNNs in supervised and semi-supervised settings, with the goal of incorporating label dependencies. Because current GNNs are not universal (i.e., most-expressive) graph representations, we propose a general collective learning approach to increase the representation power of any existing GNN. Our framework combines ideas from collective classification with self-supervised learning, and uses a Monte Carlo approach to sampling embeddings for inductive learning across graphs. We evaluate performance on five real-world network datasets and demonstrate consistent, significant improvement in node classification accuracy, for a variety of state-of-the-art GNNs.
We present a new method to learn video representations from large-scale unlabeled video data. Ideally, this representation will be generic and transferable, directly usable for new tasks such as action recognition and zero or few-shot learning. We formulate unsupervised representation learning as a multi-modal, multi-task learning problem, where the representations are shared across different modalities via distillation. Further, we introduce the concept of loss function evolution by using an evolutionary search algorithm to automatically find optimal combination of loss functions capturing many (self-supervised) tasks and modalities. Thirdly, we propose an unsupervised representation evaluation metric using distribution matching to a large unlabeled dataset as a prior constraint, based on Zipf's law. This unsupervised constraint, which is not guided by any labeling, produces similar results to weakly-supervised, task-specific ones. The proposed unsupervised representation learning results in a single RGB network and outperforms previous methods. Notably, it is also more effective than several label-based methods (e.g., ImageNet), with the exception of large, fully labeled video datasets.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191