arXiv - CS - Logic in Computer Science最新文献_第10页

A Reliable Common-Sense Reasoning Socialbot Built Using LLMs and Goal-Directed ASP 利用 LLM 和目标导向 ASP 构建可靠的常识推理社交机器人

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-26 DOI: arxiv-2407.18498

Yankai Zeng, Abhiramon Rajashekharan, Kinjal Basu, Huaduo Wang, Joaquín Arias, Gopal Gupta

{"title":"A Reliable Common-Sense Reasoning Socialbot Built Using LLMs and Goal-Directed ASP","authors":"Yankai Zeng, Abhiramon Rajashekharan, Kinjal Basu, Huaduo Wang, Joaquín Arias, Gopal Gupta","doi":"arxiv-2407.18498","DOIUrl":"https://doi.org/arxiv-2407.18498","url":null,"abstract":"The development of large language models (LLMs), such as GPT, has enabled the\u0000construction of several socialbots, like ChatGPT, that are receiving a lot of\u0000attention for their ability to simulate a human conversation. However, the\u0000conversation is not guided by a goal and is hard to control. In addition,\u0000because LLMs rely more on pattern recognition than deductive reasoning, they\u0000can give confusing answers and have difficulty integrating multiple topics into\u0000a cohesive response. These limitations often lead the LLM to deviate from the\u0000main topic to keep the conversation interesting. We propose AutoCompanion, a\u0000socialbot that uses an LLM model to translate natural language into predicates\u0000(and vice versa) and employs commonsense reasoning based on Answer Set\u0000Programming (ASP) to hold a social conversation with a human. In particular, we\u0000rely on s(CASP), a goal-directed implementation of ASP as the backend. This\u0000paper presents the framework design and how an LLM is used to parse user\u0000messages and generate a response from the s(CASP) engine output. To validate\u0000our proposal, we describe (real) conversations in which the chatbot's goal is\u0000to keep the user entertained by talking about movies and books, and s(CASP)\u0000ensures (i) correctness of answers, (ii) coherence (and precision) during the\u0000conversation, which it dynamically regulates to achieve its specific purpose,\u0000and (iii) no deviation from the main topic.","PeriodicalId":501208,"journal":{"name":"arXiv - CS - Logic in Computer Science","volume":"56 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141865893","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Repairing Networks of $mathcal{EL_perp}$ Ontologies using Weakening and Completing -- Extended version 使用弱化和补全修复 $mathcal{EL_perp}$ 本体网络 -- 扩展版

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-26 DOI: arxiv-2407.18848

Ying Li, Patrick Lambrix

引用次数: 0

Supercritical Size-Width Tree-Like Resolution Trade-Offs for Graph Isomorphism 图形同构的超临界大小-宽度树状分辨率权衡

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-25 DOI: arxiv-2407.17947

Christoph Berkholz, Moritz Lichter, Harry Vinall-Smeeth

{"title":"Supercritical Size-Width Tree-Like Resolution Trade-Offs for Graph Isomorphism","authors":"Christoph Berkholz, Moritz Lichter, Harry Vinall-Smeeth","doi":"arxiv-2407.17947","DOIUrl":"https://doi.org/arxiv-2407.17947","url":null,"abstract":"We study the refutation complexity of graph isomorphism in the tree-like\u0000resolution calculus. Tor'an and W\"orz (TOCL 2023) showed that there is a\u0000resolution refutation of narrow width $k$ for two graphs if and only if they\u0000can be distinguished in ($k+1$)-variable first-order logic (FO$^{k+1}$) and\u0000hence by a count-free variant of the $k$-dimensional Weisfeiler-Leman\u0000algorithm. While DAG-like narrow width $k$ resolution refutations have size at\u0000most $n^k$, tree-like refutations may be much larger. We show that there are\u0000graphs of order n, whose isomorphism can be refuted in narrow width $k$ but\u0000only in tree-like size $2^{Omega(n^{k/2})}$. This is a supercritical trade-off\u0000where bounding one parameter (the narrow width) causes the other parameter (the\u0000size) to grow above its worst case. The size lower bound is super-exponential\u0000in the formula size and improves a related supercritical width versus tree-like\u0000size trade-off by Razborov (JACM 2016). To prove our result, we develop a new\u0000variant of the $k$-pebble EF-game for FO$^k$ to reason about tree-like\u0000refutation size in a similar way as the Prover-Delayer games in proof\u0000complexity. We analyze this game on a modified variant of the compressed CFI\u0000graphs introduced by Grohe, Lichter, Neuen, and Schweitzer (FOCS 2023). Using a\u0000recent improved robust compressed CFI construction of Janett, Nordstr\"om, and\u0000Pang (unpublished manuscript), we obtain a similar bound for width $k$ (instead\u0000of the stronger but less common narrow width) and make the result more robust.","PeriodicalId":501208,"journal":{"name":"arXiv - CS - Logic in Computer Science","volume":"43 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141785693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

C2P: Featuring Large Language Models with Causal Reasoning C2P：具有因果推理功能的大型语言模型

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-25 DOI: arxiv-2407.18069

Abdolmahdi Bagheri, Matin Alinejad, Kevin Bello, Alireza Akhondi-Asl

{"title":"C2P: Featuring Large Language Models with Causal Reasoning","authors":"Abdolmahdi Bagheri, Matin Alinejad, Kevin Bello, Alireza Akhondi-Asl","doi":"arxiv-2407.18069","DOIUrl":"https://doi.org/arxiv-2407.18069","url":null,"abstract":"Causal reasoning is the primary bottleneck that Large Language Models (LLMs)\u0000must overcome to attain human-level intelligence. To address this, we introduce\u0000the Causal Chain of Prompting (C2P) as the first reasoning framework that\u0000equips current LLMs with causal reasoning capabilities. C2P operates\u0000autonomously, avoiding reliance on external tools or modules during both the\u0000causal learning and reasoning phases, and can be seamlessly implemented during\u0000the training or fine-tuning of LLMs. Experimental results across various\u0000benchmark datasets demonstrate a significant improvement in causal learning and\u0000subsequent reasoning accuracy of LLMs. We illustrate how C2P enhances LLMs'\u0000ability to causally reason in real-world scenarios, addressing complex problems\u0000in fields such as healthcare, medicine, economics, education, social sciences,\u0000environmental science, and marketing. With few-shot learning, GPT-4 Turbo using\u0000C2P with as few as six examples achieves significant performance improvements,\u0000boasting over a 33% increase in reasoning accuracy over the most\u0000state-of-the-art LLMs, which perform nearly randomly in similar circumstances.\u0000This demonstrates the transformative potential of integrating C2P into LLM\u0000training or fine-tuning processes, thereby empowering these models with\u0000advanced causal reasoning capabilities.","PeriodicalId":501208,"journal":{"name":"arXiv - CS - Logic in Computer Science","volume":"94 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141775998","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

On Polynomial-Time Decidability of k-Negations Fragments of First-Order Theories 论一阶理论 k 否定片段的多项式时间可决性

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-25 DOI: arxiv-2407.18420

Christoph Haase, Alessio Mansutti, Amaury Pouly

引用次数: 0

Pruning Boolean d-DNNF Circuits Through Tseitin-Awareness 通过 Tseitin-Awareness 修剪布尔 d-DNNF 电路

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-25 DOI: arxiv-2407.17951

Vincent Derkinderen

引用次数: 0

The Existential Theory of the Reals as a Complexity Class: A Compendium 作为复杂性类别的实数存在论：简编

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-25 DOI: arxiv-2407.18006

Marcus Schaefer, Jean Cardinal, Tillmann Miltzow

{"title":"The Existential Theory of the Reals as a Complexity Class: A Compendium","authors":"Marcus Schaefer, Jean Cardinal, Tillmann Miltzow","doi":"arxiv-2407.18006","DOIUrl":"https://doi.org/arxiv-2407.18006","url":null,"abstract":"We survey the complexity class $exists mathbb{R}$, which captures the\u0000complexity of deciding the existential theory of the reals. The class $exists\u0000mathbb{R}$ has roots in two different traditions, one based on the\u0000Blum-Shub-Smale model of real computation, and the other following work by\u0000Mn\"{e}v and Shor on the universality of realization spaces of oriented\u0000matroids. Over the years the number of problems for which $exists mathbb{R}$\u0000rather than NP has turned out to be the proper way of measuring their\u0000complexity has grown, particularly in the fields of computational geometry,\u0000graph drawing, game theory, and some areas in logic and algebra. $exists\u0000mathbb{R}$ has also started appearing in the context of machine learning,\u0000Markov decision processes, and probabilistic reasoning. We have aimed at collecting a comprehensive compendium of problems complete\u0000and hard for $exists mathbb{R}$, as well as a long list of open problems. The\u0000compendium is presented in the third part of our survey; a tour through the\u0000compendium and the areas it touches on makes up the second part. The first part\u0000introduces the reader to the existential theory of the reals as a complexity\u0000class, discussing its history, motivation and prospects as well as some\u0000technical aspects.","PeriodicalId":501208,"journal":{"name":"arXiv - CS - Logic in Computer Science","volume":"61 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141776002","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A quantitative probabilistic relational Hoare logic 定量概率关系胡尔逻辑

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-24 DOI: arxiv-2407.17127

Martin Avanzini, Gilles Barthe, Davide Davoli, Benjamin Grégoire

引用次数: 0

Static and Dynamic Verification of OCaml Programs: The Gospel Ecosystem (Extended Version) OCaml 程序的静态和动态验证：福音生态系统（扩展版）

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-24 DOI: arxiv-2407.17289

Tiago Lopes Soares, Ion Chririca, Mário Pereira

引用次数: 0

A process algebraic framework for multi-agent dynamic epistemic systems 多代理动态认识系统的过程代数框架

arXiv - CS - Logic in Computer Science Pub Date : 2024-07-24 DOI: arxiv-2407.17537

Alessandro Aldini

引用次数: 0