The ability to provide comprehensive explanations of chosen actions is a hallmark of intelligence. Lack of this ability impedes the general acceptance of AI and robot systems in critical tasks. This paper examines what forms of explanations best foster human trust in machines and proposes a framework in which explanations are generated from both functional and mechanistic perspectives. The robot system learns from human demonstrations to open medicine bottles using (i) an embodied haptic prediction model to extract knowledge from sensory feedback, (ii) a stochastic grammar model induced to capture the compositional structure of a multistep task, and (iii) an improved Earley parsing algorithm to jointly leverage both the haptic and grammar models. The robot system not only shows the ability to learn from human demonstrators but also succeeds in opening new, unseen bottles. Using different forms of explanations generated by the robot system, we conducted a psychological experiment to examine what forms of explanations best foster human trust in the robot. We found that comprehensive and real-time visualizations of the robot’s internal decisions were more effective in promoting human trust than explanations based on summary text descriptions. In addition, forms of explanation that are best suited to foster trust do not necessarily correspond to the model components contributing to the best task performance. This divergence shows a need for the robotics community to integrate model components to enhance both task execution and human trust in machines.
more »
« less
Montague Grammar Induction
We propose a computational modeling framework for inducing combinatory categorial grammars from arbitrary behavioral data. This framework provides the analyst fine-grained control over the assumptions that the induced grammar should conform to: (i) what the primitive types are; (ii) how complex types are constructed; (iii) what set of combinators can be used to combine types; and (iv) whether (and to what) the types of some lexical items should be fixed. In a proof-of-concept experiment, we deploy our framework for use in distributional analysis. We focus on the relationship between s(emantic)-selection and c(ategory)-selection,
using as input a lexicon-scale acceptability judgment dataset focused on English verbs’ syntactic distribution (the MegaAcceptability dataset) and enforcing standard assumptions from the semantics literature on the induced grammar.
more »
« less
- Award ID(s):
- 1940981
- NSF-PAR ID:
- 10299985
- Date Published:
- Journal Name:
- Proceedings from Semantics and Linguistic Theory
- Volume:
- 30
- ISSN:
- 2163-5951
- Page Range / eLocation ID:
- 227-251
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)We propose a computational model for inducing full-fledged combinatory categorial grammars from behavioral data. This model contrasts with prior computational models of selection in representing syntactic and semantic types as structured (rather than atomic) objects, enabling direct interpretation of the modeling results relative to standard formal frameworks. We investigate the grammar our model induces when fit to a lexicon-scale acceptability judgment dataset – Mega Acceptability – focusing in particular on the types our model assigns to clausal complements and the predicates that select them.more » « less
-
Lierler, Yuliya ; Morales, Jose F ; Dodaro, Carmine ; Dahl, Veroniica ; Gebser, Martin ; Tekle, Tuncay (Ed.)Knowledge representation and reasoning (KRR) systems represent knowledge as collections of facts and rules. Like databases, KRR systems contain information about domains of human activities like industrial enterprises, science, and business. KRRs can represent complex concepts and relations, and they can query and manipulate information in sophisticated ways. Unfortunately, the KRR technology has been hindered by the fact that specifying the requisite knowledge requires skills that most domain experts do not have, and professional knowledge engineers are hard to find. One solution could be to extract knowledge from English text, and a number of works have attempted to do so (OpenSesame, Google's Sling, etc.). Unfortunately, at present, extraction of logical facts from unrestricted natural language is still too inaccurate to be used for reasoning, while restricting the grammar of the language (so-called controlled natural language, or CNL) is hard for the users to learn and use. Nevertheless, some recent CNL-based approaches, such as the Knowledge Authoring Logic Machine (KALM), have shown to have very high accuracy compared to others, and a natural question is to what extent the CNL restrictions can be lifted. In this paper, we address this issue by transplanting the KALM framework to a neural natural language parser, mStanza. Here we limit our attention to authoring facts and queries and therefore our focus is what we call factual English statements. Authoring other types of knowledge, such as rules, will be considered in our followup work. As it turns out, neural network based parsers have problems of their own and the mistakes they make range from part-of-speech tagging to lemmatization to dependency errors. We present a number of techniques for combating these problems and test the new system, KALMFL (i.e., KALM for factual language), on a number of benchmarks, which show KALMFL achieves correctness in excess of 95%.more » « less
-
Sexual selection is a powerful force shaping not only the details but also the breadth of what we see in nature. Yet so much unexplained variation remains. Organisms often solve the “problem” of how to pass on their genes in ways that do not fit our current expectations. I argue here that integrating empirical surprises will push our understanding of sexual selection forward. Such “nonmodel” organisms (i.e., species that do not do what we think they should do) challenge us to think deeply, integrate puzzling results, question our assumptions, and consider the new (and arguably better) questions these unexpected patterns pose. In this article, I share how puzzling observations from my long-term research on the ocellated wrasse (Symphodus ocellatus) have shaped my understanding of sexual selection and suggested new questions about the interplay among sexual selection, plasticity, and social interactions. My general premise, however, is not that others should study these questions. Instead, I argue for a change in the culture of our field—to consider unexpected results a welcome opportunity to generate new questions and learn new things about sexual selection. Those of us in positions of power (e.g., as editors, reviewers, and authors) need to lead the way.more » « less
-
more » « less
Abstract Background The language of the science curriculum is complex, even in the early grades. To communicate their scientific observations, children must produce complex syntax, particularly complement clauses (e.g.,
I think it will float ;We noticed that it vibrates ). Complex syntax is often challenging for children with developmental language disorder (DLD), and thus their learning and communication of science may be compromised.Aims We asked whether recast therapy delivered in the context of a science curriculum led to gains in complement clause use and scientific content knowledge. To understand the efficacy of recast therapy, we compared changes in science and language knowledge in children who received treatment for complement clauses embedded in a first‐grade science curriculum to two active control conditions (vocabulary + science, phonological awareness + science).
Methods & Procedures This 2‐year single‐site three‐arm parallel randomized controlled trial was conducted in Delaware, USA. Children with DLD, not yet in first grade and with low accuracy on complement clauses, were eligible. Thirty‐three 4–7‐year‐old children participated in the summers of 2018 and 2019 (2020 was cancelled due to COVID‐19). We assigned participants to arms using 1:1:1 pseudo‐random allocation (avoiding placing siblings together). The intervention consisted of 39 small‐group sessions of recast therapy, robust vocabulary instruction or phonological awareness intervention during eight science units over 4 weeks, followed by two science units (1 week) taught without language intervention. Pre‐/post‐measures were collected 3 weeks before and after camp by unmasked assessors.
Outcomes & Results Primary outcome measures were accuracy on a 20‐item probe of complement clause production and performance on ten 10‐item unit tests (eight science + language, two science only). Complete data were available for 31 children (10 grammar, 21 active control); two others were lost to follow‐up. Both groups made similar gains on science unit tests for science + language content (pre versus post,
d = 2.9,p < 0.0001; group,p = 0.24). The grammar group performed significantly better at post‐test than the active control group (d = 2.5,p = 0.049) on complement clause probes and marginally better on science‐only unit tests (d = 2.5,p = 0.051).Conclusions & Implications Children with DLD can benefit from language intervention embedded in curricular content and learn both language and science targets taught simultaneously. Tentative findings suggest that treatment for grammar targets may improve academic outcomes.
What this paper adds What is already known on the subject We know that recast therapy focused on morphology is effective but very time consuming. Treatment for complex syntax in young children has preliminary efficacy data available. Prior research provides mixed evidence as to children’s ability to learn language targets in conjunction with other information.
What this study adds This study provides additional data supporting the efficacy of intensive complex syntax recast therapy for children ages 4–7 with Developmental Language Disorder. It also provides data that children can learn language targets and science curricular content simultaneously.
What are the clinical implications of this work? As SLPs, we have to talk about something to deliver language therapy; we should consider talking about curricular content. Recast therapy focused on syntactic frames is effective with young children.
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/DOI: 10.3765/salt.v30i0.4816
