arXiv subjects

Chen Jason Zhang

Publications and source records attributed to Chen Jason Zhang.

57 records · Page 4Linked to original sources

arXiv2023

Self-Convinced Prompting: Few-Shot Question Answering with Repeated Introspection

While large language models (LLMs) such as ChatGPT and PaLM have demonstrated remarkable performance in various language understanding and generation tasks, their capabilities in complex reasoning and intricate knowledge utilization still fall short of human-level proficiency. Recent studies have established the effectiveness of prompts in steering LLMs towards generating desired outputs. Building on these insights, we introduce a novel framework that harnesses the potential of large-scale pre-trained language models, to iteratively enhance performance of the LLMs. Our framework incorporates three components: \textit{Normal CoT}, a \textit{Convincer}, and an \textit{Answerer}. It processes the output of a typical few-shot chain-of-thought prompt, assesses the correctness of the response, scrutinizes the answer, refines the reasoning, and ultimately produces a new solution. Experimental results on the 7 datasets of miscellaneous problems validate the efficacy of the Self-Convince framework, achieving substantial improvements compared to the baselines. This study contributes to the burgeoning body of research focused on integrating pre-trained language models with tailored prompts and iterative refinement processes to augment their performance in complex tasks.

cs.CL↗

arXiv2018

Reducing Uncertainty of Schema Matching via Crowdsourcing with Accuracy Rates

Schema matching is a central challenge for data integration systems. Inspired by the popularity and the success of crowdsourcing platforms, we explore the use of crowdsourcing to reduce the uncertainty of schema matching. Since crowdsourcing platforms are most effective for simple questions, we assume that each Correspondence Correctness Question (CCQ) asks the crowd to decide whether a given correspondence should exist in the correct matching. Furthermore, members of a crowd may sometimes return incorrect answers with different probabilities. Accuracy rates of individual crowd workers are probabilities of returning correct answers which can be attributes of CCQs as well as evaluations of individual workers. We prove that uncertainty reduction equals to entropy of answers minus entropy of crowds and show how to obtain lower and upper bounds for it. We propose frameworks and efficient algorithms to dynamically manage the CCQs to maximize the uncertainty reduction within a limited budget of questions. We develop two novel approaches, namely `Single CCQ' and `Multiple CCQ', which adaptively select, publish and manage questions. We verify the value of our solutions with simulation and real implementation.

cs.DB↗

arXiv2017

CrowdFusion: A Crowdsourced Approach on Data Fusion Refinement

Data fusion has played an important role in data mining because high-quality data is required in a lot of applications. As on-line data may be out-of-date and errors in the data may propagate with copying and referring between sources, it is hard to achieve satisfying results with merely applying existing data fusion methods to fuse Web data. In this paper, we make use of the crowd to achieve high-quality data fusion result. We design a framework selecting a set of tasks to ask crowds in order to improve the confidence of data. Since data are correlated and crowds may provide incorrect answers, how to select a proper set of tasks to ask the crowd is a very challenging problem. In this paper, we design an approximation solution to address this challenge since we prove that the problem is at NP-hard. To further improve the efficiency, we design a pruning strategy and a preprocessing method, which effectively improve the performance of the proposed approximation solution. Furthermore, we find that under certain scenarios, we are not interested in all the facts, but only a specific set of facts. Thus, for these specific scenarios, we also develop another approximation solution which is much faster than the general approximation solution. We verify the solutions with extensive experiments on a real crowdsourcing platform.

cs.DB↗