By conditioning on natural language instructions, large language models (LLMs) have displayed impressive capabilities as general-purpose computers. However, task performance depends significantly on the quality of the prompt used to steer the model, and most effective prompts have been handcrafted by humans. Inspired by classical program synthesis and the human approach to prompt engineering, we propose Automatic Prompt Engineer (APE) for automatic instruction generation and selection. In our method, we treat the instruction as the "program," optimized by searching over a pool of instruction candidates proposed by an LLM in order to maximize a chosen score function. To evaluate the quality of the selected instruction, we evaluate the zero-shot performance of another LLM following the selected instruction. Experiments on 24 NLP tasks show that our automatically generated instructions outperform the prior LLM baseline by a large margin and achieve better or comparable performance to the instructions generated by human annotators on 19/24 tasks. We conduct extensive qualitative and quantitative analyses to explore the performance of APE. We show that APE-engineered prompts can be applied to steer models toward truthfulness and/or informativeness, as well as to improve few-shot learning performance by simply prepending them to standard in-context learning prompts. Please check out our webpage at //sites.google.com/view/automatic-prompt-engineer.
相關內容
Creating programs to correctly manipulate data is a difficult task, as the underlying programming languages and APIs can be challenging to learn for many users who are not skilled programmers. Large language models (LLMs) demonstrate remarkable potential for generating code from natural language, but in the data manipulation domain, apart from the natural language (NL) description of the intended task, we also have the dataset on which the task is to be performed, or the "data context". Existing approaches have utilized data context in a limited way by simply adding relevant information from the input data into the prompts sent to the LLM. In this work, we utilize the available input data to execute the candidate programs generated by the LLMs and gather their outputs. We introduce semantic reranking, a technique to rerank the programs generated by LLMs based on three signals coming the program outputs: (a) semantic filtering and well-formedness based score tuning: do programs even generate well-formed outputs, (b) semantic interleaving: how do the outputs from different candidates compare to each other, and (c) output-based score tuning: how do the outputs compare to outputs predicted for the same task. We provide theoretical justification for semantic interleaving. We also introduce temperature mixing, where we combine samples generated by LLMs using both high and low temperatures. We extensively evaluate our approach in three domains, namely databases (SQL), data science (Pandas) and business intelligence (Excel's Power Query M) on a variety of new and existing benchmarks. We observe substantial gains across domains, with improvements of up to 45% in top-1 accuracy and 34% in top-3 accuracy.
Scaling language models have revolutionized widespread NLP tasks, yet little comprehensively explored few-shot relation extraction with large language models. In this paper, we investigate principal methodologies, in-context learning and data generation, for few-shot relation extraction via GPT-3.5 through exhaustive experiments. To enhance few-shot performance, we further propose task-related instructions and schema-constrained data generation. We observe that in-context learning can achieve performance on par with previous prompt learning approaches, and data generation with the large language model can boost previous solutions to obtain new state-of-the-art few-shot results on four widely-studied relation extraction datasets. We hope our work can inspire future research for the capabilities of large language models in few-shot relation extraction. Code is available in \url{//github.com/zjunlp/DeepKE/tree/main/example/llm.
Large Language Models (LLMs) trained using massive text datasets have recently shown promise in generating action plans for robotic agents from high level text queries. However, these models typically do not consider the robot's environment, resulting in generated plans that may not actually be executable, due to ambiguities in the planned actions or environmental constraints. In this paper, we propose an approach to generate environmentally-aware action plans that agents are better able to execute. Our approach involves integrating environmental objects and object relations as additional inputs into LLM action plan generation to provide the system with an awareness of its surroundings, resulting in plans where each generated action is mapped to objects present in the scene. We also design a novel scoring function that, along with generating the action steps and associating them with objects, helps the system disambiguate among object instances and take into account their states. We evaluated our approach using the VirtualHome simulator and the ActivityPrograms knowledge base and found that action plans generated from our system had a 310% improvement in executability and a 147% improvement in correctness over prior work. The complete code and a demo of our method is publicly available at //github.com/hri-ironlab/scene_aware_language_planner.
Large language models (LLMs) have demonstrated remarkable potential in handling multilingual machine translation (MMT). In this paper, we systematically investigate the advantages and challenges of LLMs for MMT by answering two questions: 1) How well do LLMs perform in translating a massive number of languages? 2) Which factors affect LLMs' performance in translation? We evaluate popular LLMs, including XGLM, OPT, BLOOMZ, and ChatGPT, on 102 languages. Our empirical results show that even the best model ChatGPT still lags behind the supervised baseline NLLB in 83.33% of translation directions. Through further analysis, we discover that LLMs exhibit new working patterns when used for MMT. First, prompt semantics can surprisingly be ignored when given in-context exemplars, where LLMs still show strong performance even with unreasonable prompts. Second, cross-lingual exemplars can provide better task instruction for low-resource translation than exemplars in the same language pairs. Third, we observe the overestimated performance of BLOOMZ on dataset Flores-101, indicating the potential risk when using public datasets for evaluation.
Reasoning over knowledge graphs (KGs) is a challenging task that requires a deep understanding of the complex relationships between entities and the underlying logic of their relations. Current approaches rely on learning geometries to embed entities in vector space for logical query operations, but they suffer from subpar performance on complex queries and dataset-specific representations. In this paper, we propose a novel decoupled approach, Language-guided Abstract Reasoning over Knowledge graphs (LARK), that formulates complex KG reasoning as a combination of contextual KG search and abstract logical query reasoning, to leverage the strengths of graph extraction algorithms and large language models (LLM), respectively. Our experiments demonstrate that the proposed approach outperforms state-of-the-art KG reasoning methods on standard benchmark datasets across several logical query constructs, with significant performance gain for queries of higher complexity. Furthermore, we show that the performance of our approach improves proportionally to the increase in size of the underlying LLM, enabling the integration of the latest advancements in LLMs for logical reasoning over KGs. Our work presents a new direction for addressing the challenges of complex KG reasoning and paves the way for future research in this area.
The causal capabilities of large language models (LLMs) is a matter of significant debate, with critical implications for the use of LLMs in societally impactful domains such as medicine, science, law, and policy. We further our understanding of LLMs and their causal implications, considering the distinctions between different types of causal reasoning tasks, as well as the entangled threats of construct and measurement validity. LLM-based methods establish new state-of-the-art accuracies on multiple causal benchmarks. Algorithms based on GPT-3.5 and 4 outperform existing algorithms on a pairwise causal discovery task (97%, 13 points gain), counterfactual reasoning task (92%, 20 points gain), and actual causality (86% accuracy in determining necessary and sufficient causes in vignettes). At the same time, LLMs exhibit unpredictable failure modes and we provide some techniques to interpret their robustness. Crucially, LLMs perform these causal tasks while relying on sources of knowledge and methods distinct from and complementary to non-LLM based approaches. Specifically, LLMs bring capabilities so far understood to be restricted to humans, such as using collected knowledge to generate causal graphs or identifying background causal context from natural language. We envision LLMs to be used alongside existing causal methods, as a proxy for human domain knowledge and to reduce human effort in setting up a causal analysis, one of the biggest impediments to the widespread adoption of causal methods. We also see existing causal methods as promising tools for LLMs to formalize, validate, and communicate their reasoning especially in high-stakes scenarios. In capturing common sense and domain knowledge about causal mechanisms and supporting translation between natural language and formal methods, LLMs open new frontiers for advancing the research, practice, and adoption of causality.
The field of machine learning (ML) has gained widespread adoption, leading to a significant demand for adapting ML to specific scenarios, which is yet expensive and non-trivial. The predominant approaches towards the automation of solving ML tasks (e.g., AutoML) are often time consuming and hard to understand for human developers. In contrast, though human engineers have the incredible ability to understand tasks and reason about solutions, their experience and knowledge are often sparse and difficult to utilize by quantitative approaches. In this paper, we aim to bridge the gap between machine intelligence and human knowledge by introducing a novel framework MLCopilot, which leverages the state-of-the-art LLMs to develop ML solutions for novel tasks. We showcase the possibility of extending the capability of LLMs to comprehend structured inputs and perform thorough reasoning for solving novel ML tasks. And we find that, after some dedicated design, the LLM can (i) observe from the existing experiences of ML tasks and (ii) reason effectively to deliver promising results for new tasks. The solution generated can be used directly to achieve high levels of competitiveness.
Remote photoplethysmography (rPPG) offers a state-of-the-art, non-contact methodology for estimating human pulse by analyzing facial videos. Despite its potential, rPPG methods can be susceptible to various artifacts, such as noise, occlusions, and other obstructions caused by sunglasses, masks, or even involuntary facial contact, such as individuals inadvertently touching their faces. In this study, we apply image processing transformations to intentionally degrade video quality, mimicking these challenging conditions, and subsequently evaluate the performance of both non-learning and learning-based rPPG methods on the deteriorated data. Our results reveal a significant decrease in accuracy in the presence of these artifacts, prompting us to propose the application of restoration techniques, such as denoising and inpainting, to improve heart-rate estimation outcomes. By addressing these challenging conditions and occlusion artifacts, our approach aims to make rPPG methods more robust and adaptable to real-world situations. To assess the effectiveness of our proposed methods, we undertake comprehensive experiments on three publicly available datasets, encompassing a wide range of scenarios and artifact types. Our findings underscore the potential to construct a robust rPPG system by employing an optimal combination of restoration algorithms and rPPG techniques. Moreover, our study contributes to the advancement of privacy-conscious rPPG methodologies, thereby bolstering the overall utility and impact of this innovative technology in the field of remote heart-rate estimation under realistic and diverse conditions.
Large language models (LLMs), such as Codex, hold great promise in enhancing programming education by automatically generating feedback for students. We investigate using LLMs to generate feedback for fixing syntax errors in Python programs, a key scenario in introductory programming. More concretely, given a student's buggy program, our goal is to generate feedback comprising a fixed program along with a natural language explanation describing the errors/fixes, inspired by how a human tutor would give feedback. While using LLMs is promising, the critical challenge is to ensure high precision in the generated feedback, which is imperative before deploying such technology in classrooms. The main research question we study is: Can we develop LLMs-based feedback generation techniques with a tunable precision parameter, giving educators quality control over the feedback that students receive? To this end, we introduce PyFiXV, our technique to generate high-precision feedback powered by Codex. The key idea behind PyFiXV is to use a novel run-time validation mechanism to decide whether the generated feedback is suitable for sharing with the student; notably, this validation mechanism also provides a precision knob to educators. We perform an extensive evaluation using two real-world datasets of Python programs with syntax errors and show the efficacy of PyFiXV in generating high-precision feedback.
To make progress towards multi-modal AI assistants which can guide users to achieve complex multi-step goals, we propose the task of Visual Planning for Assistance (VPA). Given a goal briefly described in natural language, e.g., "make a shelf", and a video of the user's progress so far, the aim of VPA is to obtain a plan, i.e., a sequence of actions such as "sand shelf", "paint shelf", etc., to achieve the goal. This requires assessing the user's progress from the untrimmed video, and relating it to the requirements of underlying goal, i.e., relevance of actions and ordering dependencies amongst them. Consequently, this requires handling long video history, and arbitrarily complex action dependencies. To address these challenges, we decompose VPA into video action segmentation and forecasting. We formulate the forecasting step as a multi-modal sequence modeling problem and present Visual Language Model based Planner (VLaMP), which leverages pre-trained LMs as the sequence model. We demonstrate that VLaMP performs significantly better than baselines w.r.t all metrics that evaluate the generated plan. Moreover, through extensive ablations, we also isolate the value of language pre-training, visual observations, and goal information on the performance. We will release our data, model, and code to enable future research on visual planning for assistance.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191