Do large language models solve verbal analogies like children do?
Analogical reasoning –learning about new things by relating it to previous knowledge– lies at the heart of human intelligence and creativity and forms the core of educational practice. Children start creating and using analogies early on, making incredible progress moving from associative processes to successful analogical reasoning. For example, if we ask a four-year-old “Horse belongs to stable like chicken belongs to …?” they may use association and reply “egg”, whereas older children will likely give the intended relational response “chicken coop” (or other term to refer to a chicken’s home). Interestingly, despite state-of-the-art AI-language models having superhuman encyclopedic knowledge and superior memory and computational power, our pilot studies show that these large language models often make mistakes providing associative rather than relational responses to verbal analogies. For example, when we asked four- to eight-year-olds to solve the analogy “body is to feet as tree is to …?” they responded “roots” without hesitation, but large language models tend to provide more associative responses such as “leaves”. In this study we examine the similarities and differences between children's and six large language models' (Dutch/multilingual models: RobBERT, BERT-je, M-BERT, GPT-2, M-GPT, Word2Vec and Fasttext) responses to verbal analogies extracted from an online adaptive learning environment, where >14,000 7-12 year-olds from the Netherlands solved 20 or more items from a database of 900 Dutch language verbal analogies.
From the Didactic to the Heuristic Use of Analogies in Science Teaching
Extensive research on science teaching has shown the effectiveness of analogies as a didactic tool which, when appropriately and effectively used, facilitates the learning process of abstract concepts. This seminar does not contradict the efficacy of such a didactic use of analogies in this seminar but switches attention and interest on their heuristic use in approaching and understanding of what previously unknown. Such a use of analogies derives from research with 10 to 17 year-olds, who, when asked to make predictions in novel situations and to then provide explanations about these predictions, they self-generated analogies and used them by reasoning on their basis. This heuristic use of analogies can be used in science teaching in revealing how students approach situations they have not considered before as well as the sources they draw upon in doing so.
Reasoning Ability: Neural Mechanisms, Development, and Plasticity
Relational thinking, or the process of identifying and integrating relations between mental representations, is regularly invoked during reasoning. This mental capacity enables us to draw higher-order abstractions and generalize across situations and contexts, and we have argued that it should be included in the pantheon of executive functions. In this talk, I will briefly review our lab's work characterizing the roles of lateral prefrontal and parietal regions in relational thinking. I will then discuss structural and functional predictors of individual differences and developmental changes in reasoning.
The Limits of Causal Reasoning in Human and Machine Learning
A key purpose of causal reasoning by individuals and by collectives is to enhance action, to give humans yet more control over their environment. As a result, causal reasoning serves as the infrastructure of both thought and discourse. Humans represent causal systems accurately in some ways, but also show some systematic biases (we tend to neglect causal pathways other than the one we are thinking about). Even when accurate, people’s understanding of causal systems tends to be superficial; we depend on our communities for most of our causal knowledge and reasoning. Nevertheless, we are better causal reasoners than machines. Modern machine learners do not come close to matching human abilities.
Causal Reasoning: Its role in the architecture and development of the mind
The seminar will first outline the architecture of the human mind, specifying general and domain-specific mental processes. The place of causal reasoning and its relations with the other processes will be specified. Experimental, psychometric, developmental, and brain-based evidence will be summarized. The main message of the talk is that causal thought involves domain-specific core processes rooted in perception and served by special brain networks which capture interactions between objects. With development, causal reasoning is increasingly associated with a general abstraction system which generates general principles underlying inductive, analogical, and deductive reasoning and also heuristics for specifying causal relations. These associations are discussed in some detail. Possible implications for artificial intelligence and educational implications are also discussed.
Seeing things clearly: Image understanding through hard-attention and reasoning with structured knowledges
In this talk, Jonathan aims to frame the current challenges of explainability and understanding in ML-driven approaches to image processing, and their potential solution through explicit inference techniques.
Differential working memory functioning
The integrated conflict monitoring theory of Botvinick introduced cognitive demand into conflict monitoring research. We investigated effects of individual differences of cognitive demand and another determinant of conflict monitoring entitled reinforcement sensitivity on conflict monitoring. We showed evidence of differential variability of conflict monitoring intensity using the electroencephalogram (EEG), functional magnet resonance imaging (fMRI) and behavioral data. Our data suggest that individual differences of anxiety and reasoning ability are differentially related to the recruitment of proactive and reactive cognitive control (cf. Braver). Based on previous findings, the team of the Leue-Lab investigated new psychometric data on conflict monitoring and proactive-reactive cognitive control. Moreover, data of the Leue-Lab suggest the relevance of individual differences of conflict monitoring for the context of deception. In this respect, we plan new studies highlighting individual differences of the functioning of the Anterior Cingulate Cortex (ACC). Disentangling the role of individual differences in working memory-related cognitive demand, mental effort, and reinforcement-related processes opens new insights for cognitive-motivational approaches of information processing (Passcode to rewatch: 0R8v&m59).
Zero-shot visual reasoning with probabilistic analogical mapping
There has been a recent surge of interest in the question of whether and how deep learning algorithms might be capable of abstract reasoning, much of which has centered around datasets based on Raven’s Progressive Matrices (RPM), a visual analogy problem set commonly employed to assess fluid intelligence. This has led to the development of algorithms that are capable of solving RPM-like problems directly from pixel-level inputs. However, these algorithms require extensive direct training on analogy problems, and typically generalize poorly to novel problem types. This is in stark contrast to human reasoners, who are capable of solving RPM and other analogy problems zero-shot — that is, with no direct training on those problems. Indeed, it’s this capacity for zero-shot reasoning about novel problem types, i.e. fluid intelligence, that RPM was originally designed to measure. I will present some results from our recent efforts to model this capacity for zero-shot reasoning, based on an extension of a recently proposed approach to analogical mapping we refer to as Probabilistic Analogical Mapping (PAM). Our RPM model uses deep learning to extract attributed graph representations from pixel-level inputs, and then performs alignment of objects between source and target analogs using gradient descent to optimize a graph-matching objective. This extended version of PAM features a number of new capabilities that underscore the flexibility of the overall approach, including 1) the capacity to discover solutions that emphasize either object similarity or relation similarity, based on the demands of a given problem, 2) the ability to extract a schema representing the overall abstract pattern that characterizes a problem, and 3) the ability to directly infer the answer to a problem, rather than relying on a set of possible answer choices. This work suggests that PAM is a promising framework for modeling human zero-shot reasoning.
Probabilistic Analogical Mapping with Semantic Relation Networks
Hongjing Lu will present a new computational model of Probabilistic Analogical Mapping (PAM, in collaboration with Nick Ichien and Keith Holyoak) that finds systematic correspondences between inputs generated by machine learning. The model adopts a Bayesian framework for probabilistic graph matching, operating on semantic relation networks constructed from distributed representations of individual concepts (word embeddings created by Word2vec) and of relations between concepts (created by our BART model). We have used PAM to simulate a broad range of phenomena involving analogical mapping by both adults and children. Our approach demonstrates that human-like analogical mapping can emerge from comparison mechanisms applied to rich semantic representations of individual concepts and relations. More details can be found https://arxiv.org/ftp/arxiv/papers/2103/2103.16704.pdf
Analogy as a Catalyst for Cumulative Cultural Evolution
Analogies, broadly defined, map novel concepts onto familiar concepts, making them essential for perception, reasoning, and communication. We argue that analogy-building served a critical role in the evolution of cumulative culture, by allowing humans to learn and transmit complex behavioural sequences that would otherwise be too cognitively demanding or opaque to acquire. The emergence of a protolanguage consisting of simple labels would have provided early humans with the cognitive tools to build explicit analogies and to communicate them to others. This focus on analogy-building can shed new light on the coevolution of cognition and culture, and addresses recent calls for better integration of the field of cultural evolution with cognitive science. This talk will address what cumulative cultural evolution is, how we define analogy-building, how analogy-building applies to cumulative cultural evolution, how analogy-building fits into language evolution, and the implications of analogy-building for causal understanding and cognitive evolution.
Logical Neural Networks
The work to be presented in this talk proposes a novel framework seamlessly providing key properties of both neural nets (learning) and symbolic logic (knowledge and reasoning). Every neuron has a meaning as a component of a formula in a weighted real-valued logic, yielding a highly interpretable disentangled representation. Inference is omnidirectional rather than focused on predefined target variables, and corresponds to logical reasoning, including classical first-order logic theorem proving as a special case. The model is end-to-end differentiable, and learning minimizes a novel loss function capturing logical contradiction, yielding resilience to inconsistent knowledge. It also enables the open-world assumption by maintaining bounds on truth values which can have probabilistic semantics, yielding resilience to incomplete knowledge.
Abstract Semantic Relations in Mind, Brain, and Machines
Abstract semantic relations (e.g., category membership, part-whole, antonymy, cause-effect) are central to human intelligence, underlying the distinctively human ability to reason by analogy. I will describe a computational project (Bayesian Analogy with Relational Transformations) that aims to extract explicit representations of abstract semantic relations from non-relational inputs automatically generated by machine learning. BART’s representations predict patterns of typicality and similarity for semantic relations, as well as similarity of neural signals triggered by semantic relations during analogical reasoning. In this approach, analogy emerges from the ability to learn and compare relations; mapping emerges later from the ability to compare patterns of relations.