An official website of the United States government
*Uncertainty expressions* such as ‘probably’ or ‘highly unlikely’ are pervasive in human language. While prior work has established that there is population-level agreement in terms of how humans quantitatively interpret these expressions, there has been little inquiry into the abilities of language models in the same context. In this paper, we investigate how language models map linguistic expressions of uncertainty to numerical responses. Our approach assesses whether language models can employ theory of mind in this setting: understanding the uncertainty of another agent about a particular statement, independently of the model’s own certainty about that statement. We find that 7 out of 10 models are able to map uncertainty expressions to probabilistic responses in a human-like manner. However, we observe systematically different behavior depending on whether a statement is actually true or false. This sensitivity indicates that language models are substantially more susceptible to bias based on their prior knowledge (as compared to humans). These findings raise important questions and have broad implications for human-AI and AI-AI communication.
more »
« less
Leveraging Unstructured Statistical Knowledge in a Probabilistic Language of Thought
One hallmark of human reasoning is that we can bring to bear a diverse web of common-sense knowledge in any situation. The vastness of our knowledge poses a challenge for the practical implementation of reasoning systems as well as for our cognitive theories – how do people represent their common-sense knowledge? On the one hand, our best models of sophisticated reasoning are top-down, making use primarily of symbolically-encoded knowledge. On the other, much of our understanding of the statistical properties of our environment may arise in a bottom-up fashion, for example through asso- ciationist learning mechanisms. Indeed, recent advances in AI have enabled the development of billion-parameter language models that can scour for patterns in gigabytes of text from the web, picking up a surprising amount of common-sense knowledge along the way—but they fail to learn the structure of coherent reasoning. We propose combining these approaches, by embedding language-model-backed primitives into a state- of-the-art probabilistic programming language (PPL). On two open-ended reasoning tasks, we show that our PPL models with neural knowledge components characterize the distribution of human responses more accurately than the neural language models alone, raising interesting questions about how people might use language as an interface to common-sense knowledge, and suggesting that building probabilistic models with neural language-model components may be a promising approach for more human-like AI.
more »
« less
- Award ID(s):
- 1911790
- PAR ID:
- 10159026
- Date Published:
- Journal Name:
- Proceedings of the Annual Conference of the Cognitive Science Society
- ISSN:
- 1069-7977
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Current AI systems lack several important human capabilities, such as adaptability, generalizability, selfcontrol, consistency, common sense, and causal reasoning. We believe that existing cognitive theories of human decision making, such as the thinking fast and slow theory, can provide insights on how to advance AI systems towards some of these capabilities. In this paper, we propose a general architecture that is based on fast/slow solvers and a metacognitive component. We then present experimental results on the behavior of an instance of this architecture, for AI systems that make decisions about navigating in a constrained environment. We show how combining the fast and slow decision modalities, which can be implemented by learning and reasoning components respectively, allows the system to evolve over time and gradually pass from slow to fast thinking with enough experience, and that this greatly helps in decision quality, resource consumption, and efficiency.more » « less
-
CAIMIRA discovers the skills that humans and AIs use to answer questions. By scraping websites where trivia nerds answer really difficult questions and posing those questions to AI models like GPT-4 and LLaMA-3-70B, while humans excel in knowledge-based abductive reasoning, AI outperforms on fact-based historical recall. This research suggests future challenges should focus on more complex reasoning and nuanced language tasks to better align AI development with human cognitive strengths.more » « less
-
Effective teamwork depends on teammates’ ability to maintain common ground: mutual knowledge about the relevant state of the world and the relevant status of teammates’ actions and plans. This ability integrates diverse skills of reasoning and communication: agents can track common ground by recognizing and registering public updates to ongoing activity, but when this evidence is incomplete, agents may need to describe what they are doing or ask what others are doing. In this paper, we introduce an architecture for integrating these diverse skills to maintain common ground in human–AI teamwork. Our approach offers unique advantages of simplicity, modularity, and extensibility by leveraging generic tools for plan recognition, planning, natural language understanding and generation, and dialogue management. Worked examples illustrate how linguistic and practical reasoning complement each other in the realization of key interactive skills.more » « less
-
Growing up with bedtime tales, even children could easily tell how a story should develop; but selecting a coherent and reasonable ending for a story is still not easy for machines. To successfully choose an ending requires not only detailed analysis of the context, but also applying commonsense reasoning and basic knowledge. Previous work has shown that language models trained on very large corpora could capture common sense in an implicit and hard-to-interpret way. We explore another direction and present a novel method that explicitly incorporates commonsense knowledge from a structured dataset, and demonstrate the potential for improving story completion.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- The DOI is not currently available.
