Inference-Time Scaling has been critical to the success of recent models such as OpenAI o1 and DeepSeek R1. However, many techniques used to train models for inference-time scaling require tasks to have answers that c...
详细信息
Probabilistic neurosymbolic learning seeks to integrate neural networks with symbolic programming. Many state-of-the-art systems rely on a reduction to the Probabilistic Weighted Model Counting Problem (PWMC), which r...
详细信息
Large Language Models (LLMs) trained on petabytes of data are highly compressed repositories of a significant proportion of the knowledge accumulated and distilled so far. In this paper we study techniques to elicit t...
详细信息
Current scientific research witnesses various attempts at applying Large Language Models for scenario generation but is inclined only to comprehensive or dangerous scenarios. In this paper, we seek to build a three-st...
详细信息
Traditionally, linters are code analysis tools that help developers by flagging potential issues from syntax and logic errors to enforcing syntactical and stylistic conventions. Recently, linting has been taken as an ...
详细信息
One of the fundamental challenges in reinforcement learning RL is to take a complex task and be able to decompose it to subtasks that are simpler for the RL agent to learn. In this paper, we report on our work that wo...
详细信息
ISBN:
(数字)9798350368741
ISBN:
(纸本)9798350368758
One of the fundamental challenges in reinforcement learning RL is to take a complex task and be able to decompose it to subtasks that are simpler for the RL agent to learn. In this paper, we report on our work that would identify subtasks by using some given positive and negative trajectories for solving the complex task. We assume that the states are represented by first-order predicate logic using which we devise a novel algorithm to identify the subtasks. Then we employ a Large Language Model (LLM) to generate first-order logic rule templates for achieving each subtask. Such rules were then further fined tuned to a rule-based policy via an Inductive logic programming (ILP)-based RL agent. Through experiments, we verify the accuracy of our algorithm in detecting subtasks which successfully detect all of the subtasks correctly. We also investigated the quality of the common-sense rules produced by the language model to achieve the subtasks. Our experiments show that our LLM-guided rule template generation can produce rules that are necessary for solving a subtask, which leads to solving complex tasks with fewer assumptions about predefined first-order logic predicates of the environment.
The ABductive Learning (ABL) framework aims to bridge the perception and reasoning capabilities of artificial intelligence (AI) by unifying machine learning and logic programming. While the machine learning component ...
详细信息
ISBN:
(数字)9798350379037
ISBN:
(纸本)9798350379044
The ABductive Learning (ABL) framework aims to bridge the perception and reasoning capabilities of artificial intelligence (AI) by unifying machine learning and logic programming. While the machine learning component classifies symbolic labels from datasets, the logic programming aspect reasons with these labels using a knowledge base, correcting misclassifications. However, the original ABL framework relies on classical logic, which inadequately handles inconsistent information, a common occurrence in knowledge bases. This paper introduces an initial integration of paraconsistent logic programming with abductive learning, called Paraconsistent ABductive Learning (PABL), to enable reasoning among inconsistent information. An experiment on the MNIST single-digit addition task illustrates our approach, showing that our ABL extension maintains a state-of-the-art accuracy of 98.1%. The implementation of our proposed model is publicly available at https://***/LiuBodan/PABL.
Abstraction emerges as a valuable method across diverse domains of Artificial Intelligence (AI), particularly in the field of knowledge representation and reasoning. Intuitively, abstraction maps a complicated structu...
详细信息
ISBN:
(纸本)9798400704864
Abstraction emerges as a valuable method across diverse domains of Artificial Intelligence (AI), particularly in the field of knowledge representation and reasoning. Intuitively, abstraction maps a complicated structure to a simpler version of it. That reduces the computational complexity of the task being considered, as it provides us with the ability to focus on the parts of the problem that are relevant to the solution. In our view, such a tool can also have potential in the field of non-monotonic reasoning. Non-monotonicity is a crucial notion as it is very common when reasoning over defeasible knowledge. Adding new entries to our current knowledge, oftentimes results in restricting the conclusions that we can draw. For this form of reasoning we use certain formalisms, such as computational argumentation and logic programming (LP), that help us capture non-monotonicity. However, interpreting these formalisms faces hardships due to the large structures that might occur when representing the problem in question. Hence, coming up with ways to manage these structures easier is necessary. Recently, abstraction was shown to be a promising tool when dealing with Argumentation Frameworks (AFs) as well as with LP. AFs are frameworks with graph-like structure, whose nodes represent arguments with no internal structure, while edges stand for conflicts among the arguments. In our research we focus on continuing in this direction by employing structured frameworks such as Assumption-Based Argumentation Frameworks (ABAFs). Subsequently, we will extend our research to similar formalisms such as LP.
Answer set programming (ASP) is a logic programming formalism used in various areas of artificial intelligence like combinatorial problem solving and knowledge representation and reasoning. It is known that enhancing ...
详细信息
The technique of forgetting in knowledge representation has been shown to be a powerful and useful knowledge engineering tool with widespread application. Yet, very little research has been done on how different polic...
详细信息
暂无评论