United States

Knowing that something is unknown is an important part of human cognition. While Bayesian models of cognition have been successful in explaining many aspects of human learning, current explanations of how humans realise that they need to introduce a new concept are problematic. Bayesian models lack a principled way to describe ignorance, as doing so requires comparing the probabilities of concepts in the model with the probabilities of concepts not present in the model, which is by definition impossible. Formal definitions of uncertainty (e.g. Shannon-entropy) are commonly used as a substitute for ignorance, but we will show that these concepts are fundamentally distinct, and thus that something more is needed. Enhancing probability theory to allow Bayesian agents to conclude that they are ignorant would be an important advance for both cognitive engineering and cognitive science. In this research project, we formally analyse this challenge.

CogSci 2021

Certainly Strange: A Probabilistic Perspective on Ignorance

The AGI conference series is the only major conference series devoted wholly and specifically to the creation of AI systems possessing general intelligence at the human level and ultimately beyond. By gathering together active researchers in the field, for presentation of results and discussion of ideas, we accelerate our progress toward our common goal.

Underline hosted the two main days of AGI 2020 on Wednesday June 24th and Thursday June 25th. Both days were **live-streamed** and you can read the full program bellow. All presentations are available for a **asynchronous viewing**. All presenters are grouped into panels. You can also hear guest talks by **Ben Goertzel** (SingularityNET), **Valery Tarasov** (Semenov Institute of Chemical Physics), **Tatiana Shavrina** (Sberbank) and **Joscha Bach** (MIT Media Lab). 

AGI-20

The AGI conference series is the only major conference series devoted wholly and specifically to the creation of AI systems possessing general intelligence at the human level and ultimately beyond. By gathering together active researchers in the field, for presentation of results and discussion of ideas, we accelerate our progress toward our common goal.

poster

### Looking for Cogsci 2023? Please [click here](https://underline.io/events/405/reception) to switch to 2023 edition.


**Dear Colleagues,**

Welcome to the 43rd Annual Conference of the Cognitive Science Society! 

This year’s conference brings together the latest research from all Cognitive Science and also highlights this year’s theme **Comparative Cognition: Animal Minds.** 

The program features four plenary speakers: Friederike Range (University of Veterinary Medicine, Vienna), Onur Güntürkün (Ruhr University Bochum), Julia Fischer (University of Göttingen), and Erich Jarvis (The Rockefeller University). The keynotes and invited symposia showcase the broad spectrum of ideas that encompass the conference theme of comparative cognition. 

Further, the program includes the Rumelhart Prize keynote presentation by Susan Goldin-Meadow (University of Chicago), and symposia to celebrate the winner of the Elman Prize: Wei Ji Ma (New York University), the Heineken Award winner Robert Zatorre (McGill University), and the Robert J. Glushko Dissertation Prizes session.

The program committee for CogSci 2021 received 1,025 submissions, including 710 full papers, 275 member abstracts, 8 publication-based talks, as well as 15 proposals for symposia, and 17 for workshops and tutorials. After a rigorous review process, the committee selected 152 papers for oral presentation and inclusion in the conference proceedings (14.8%), 332 papers for poster presentation and inclusion in the proceedings (32.4%), and 175 papers for poster presentation with inclusion of abstracts in the proceedings (17.1%) We also selected 275 submitted member abstracts. In addition, we accepted 6 publication-based talks, 12 symposia, and 9 workshops and tutorials to make for an exciting and inclusive program.

We hope that you enjoy the conference this year!<br><br>

**Your Program Chairs,**
![](https://assets.underline.io/uploads/markdown_image/1/image/99984908caba5e6453e5fc5ba34f3507.jpg)

**Learn more about the Vienna Cognitive Science Hub at the University of Vienna, hosts of CogSci 2021**

<iframe width="560" height="315" src="https://www.youtube.com/embed/464h1m3iU68" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

To review the proceedings for the conference or review the searchable program, please visit the [website](https://cognitivesciencesociety.org/program/).


The Cognitive Science Society brings together researchers from around the world who hold a common goal: understanding the nature of the human mind. The mission of the Society is to promote Cognitive Science as a discipline, and to foster scientific interchange among researchers in various areas of study, including Artificial Intelligence, Linguistics, Anthropology, Psychology, Neuroscience, Philosophy, and Education. The Society is a non-profit professional organization and its activities include sponsoring an annual conference and publishing the journals Cognitive Science and TopiCS.

Teachers often use drawings of the normal distribution to support explanations of related statistical concepts, assuming that the normal curve provides a common language for such discussions. However, we find that students may not understand the basic features of the normal curve. In Study 1, we showed that students who already have studied the normal distribution in a college-level class do not understand basic concepts associated with it. Then, in Study 2, we investigated whether a brief instructional, narrated video could improve students’ understanding of the normal probability distribution. Specifically, we compared three instructional formats: static slides, a video recording of a hand physically drawing those plots, and a screen recording of the hand-drawing. Despite the brevity of the intervention, we found significant improvements in students’ understanding of the normal probability distribution and related probability concepts. The findings are discussed in relation to the dynamic representation and embodied cognition literature.

Utilizing Dynamic and Embodied Visualization to Facilitate Understanding of Normal Probability Distributions

Not all categories are made the same. Some categories have high within-category variability (e.g., “vehicles” can look very different) and some have low within-category variability (e.g., “apples” are pretty similar). Categories can also vary on their between-category variability where some categories are very similar to each other (e.g., “apples” and “oranges”) and some are very different (e.g., “apples” and “vehicles”). Studies have found that categories with high within and between variability are learned best in massed formats, and categories with low within and between variability are learned best in interleaved formats. However, the unique contribution of each of these kinds of variability (i.e., within and between) have not been studied independently. These studies investigate the unique contribution of within- and between-category variability to 3-year-old children’s word learning in interleaved and massed presentations. The results inform existing understanding of interleaving in word learning and how category variability impacts learning.

Effects on word learning from spacing and category variability

Mathematics, by its very nature, is rife with patterns. However, students frequently treat mathematics as a series of isolated and unlinked exercises. The current study examined whether experience with extending mathematical patterns affected adults’ ability to solve equations that involved patterns and/or to reason about mathematical relationships in new contexts. Participants who were given 13 trials of pattern extension experience then went on to demonstrate both more efficient problem-solving (i.e., faster response times) and more accurate problem-solving at posttest, relative to individuals who were given an equivalent amount of explicit instruction related to solving the equations. However, there was no difference between the groups in the ability to abstract the structure of the underlying mathematical relationships. These findings suggest that patterning tasks like those used in this study may be useful in supporting math performance.

Experience with Equations in Sequence Promotes Procedural Fluency

Visuospatial cognitive skills are increasingly recognized as critical in many areas of both formal and informal learning, but far less research has looked at their role in the workforce. We know, for example, that strong visuospatial skills in high school are predictive of STEM occupations, but we know less about how these skills are actually deployed in the workplace, and also the extent to which these skills might also be important in various non-STEM occupational pathways. Our review provides current context for these questions, including identifying occupations for which these research questions are actively being explored, as well as areas that urgently call for more research. We argue that systematic and rigorous cognitive research should be used to inform evidence-based best practices for addressing the unique workforce-related challenges of the 21st century.

Visuospatial Skills and the Workforce: A Brief Review

What visualization strategies do people use to communicate abstract knowledge to others? We developed a drawing paradigm to elicit visual explanations about novel machines and obtained detailed annotations of the semantic information conveyed in each drawing. We found that these visual explanations contained: (1) greater emphasis on causally relevant parts of the machine, (2) less emphasis on structural features that were visually salient but causally irrelevant, and (3) more symbols, relative to baseline drawings intended only to communicate the machines' appearance. However, this overall pattern of emphasis did not necessarily improve naive viewers' ability to infer how to operate the machines, nor their ability to identify them, suggesting a potential mismatch between what people believe a visual explanation contains and what may be most useful. Taken together, our findings advance our understanding of how communicative goals constrain visual communication of abstract knowledge across behavioral contexts.

How do the semantic properties of visual explanations guide causal inference?

One problem language learners face is extracting word meanings from scenes with many possible referents. Despite the ambiguity of individual situations, a large body of empirical work shows that people are able to learn cross-situationally when a word occurs in different situations. Many computational models of cross-situational word learning have been proposed, yet there is little consensus on the main mechanisms supporting learning, in part due to the profusion of disparate studies and models, and lack of systematic model comparisons across a wide range of studies. This study compares the performance of several extant models on a dataset of 44 experimental conditions and a total of 1,696 participants. Using cross-validation, we fit multiple models representing theories of both
associative learning and hypothesis-testing theories of word learning, find two best-fitting models, and discuss issues of model and mechanism identifiability. Finally, we test the models' ability to generalize to additional experiments, including developmental data.

A large-scale comparison of cross-situational word learning models

Test anxiety affects a sizable proportion of college students, especially in competitive STEM fields. Prior research has proposed interventions aimed at changing students’ implicit beliefs about intelligence (intelligence mindset) to help reduce students’ test anxiety, but results have been mixed and the mechanisms underlying this relationship are unclear. We propose that students’ beliefs about effort regulation may partially mediate the relationship between intelligence mindsets and test anxiety. We tested our model as an exploratory post-hoc analysis in a small sample of introductory physics students reporting psychological threat in a laboratory study. Effort regulation was measured as self-reported judgements of persistence in the face of difficulty and test anxiety was measured on a problem by problem basis during a laboratory physics assessment. Students’ intelligence mindset at pretest was a significant predictor of test anxiety at posttest, and this relationship was mediated by self-reported effort regulation. We discuss potential implications of these findings for mindset-based interventions aimed at reducing test anxiety.

The Relationship Between Intelligence Mindset and Test Anxiety as Mediated by Effort Regulation

Memories of most stimuli in the auditory and other domains are prone to the disruptive interference of intervening events, whereby memory performance continuously declines as the number of intervening events increases. However, melodies in a familiar musical idiom are robust to such interference. We propose that representations of musical structure emerging from syntactic processing may provide partially redundant information that accounts for this robust encoding in memory. The present study employs tonally ambiguous melodies which afford two different syntactic interpretations in the tonal idiom. Crucially, since the melodies are ambiguous, memory across two presentations of the same melody cannot bias whether the interpretation in a second listening will be the same as the first, unless a representation of the first syntactic interpretation is also encoded in memory in addition to sensory information. The melodies were presented in a Memory Task, based on a continuous recognition paradigm, as well as in a Structure Task, where participants reported their syntactic interpretation of each melody following a disambiguating cue. Our results replicate memory-for-melody's robustness to interference, and further establish a predictive relationship between memory performance in the Memory Task and the robustness of syntactic interpretations against the bias introduced by the disambiguating cue in the Structure Task. As a consequence, our results support that a representation based on a disambiguating syntactic parse provides an additional, partially redundant encoding that feeds into memory alongside sensory information. Furthermore, establishing a relationship between memory performance and the formation of structural representations supports the relevance of syntactic relationships towards the experience of music.

Musical syntactic structure improves memory for melody: evidence from the processing of ambiguous melodies

Maintaining attention during instruction is challenging as children face various sources of distraction (peers, announcements, noise) as well as competition from the visual environment itself. Prior studies found decorated environments promote off-task behavior and reduce learning. Additionally, many classroom displays are not relevant to ongoing instruction. This raises the possibility that increasing alignment between displays and instructional content may afford opportunities for incidental learning, reducing the detrimental effects of environmental off-task behavior. To investigate this possibility, participants completed a lesson in which alignment between the lesson content and displays was manipulated (relevant, educational but irrelevant, or no displays). Attention to the lesson and learning gains for content presented in the lesson and/or displays were measured. Results suggest younger children’s learning can benefit from displays that reinforce the lesson content. However, there was no evidence of incidental learning from displays without additional lesson support. Implications for classroom design are discussed.

The Decorated Learning Environment: Simply Noise or an Opportunity for Incidental Learning?

Gesture during math instruction supports learning in children and adults. The mechanism by which gesture enhances learning across development is not known. One possibility is that instruction with gesture engages different cognitive abilities during learning than instruction without gesture. Our previous work showed a positive relationship between visuospatial working memory capacity and learning only when gesture was present, and a positive relationship between verbal working memory capacity and learning only when gesture was absent, suggesting that gesture may be processed using visuospatial working memory. The aim of the current experiment was to replicate and extend these prior findings with new instruction, random assignment to instructional condition, and improved measures of both learning and cognitive abilities. Participants observed video instruction in a novel mathematical system that either included speech and gesture or only speech. After instruction, participants completed a posttest to assess learning. Finally, participants completed tasks to assess verbal and visuospatial working memory capacity as well as fluid and crystallized intelligence. We found that gesture benefitted learning in adults. Contrary to previous findings, both learning with gesture and learning without gesture were supported by visuospatial working memory. These findings suggest that changing characteristics of instruction does not necessarily change the cognitive resources supporting learning in a novel math task.

Certainly Strange: A Probabilistic Perspective on Ignorance

Downloads

Next from CogSci 2021

Utilizing Dynamic and Embodied Visualization to Facilitate Understanding of Normal Probability Distributions

Similar lecture

Towards Dynamic Process Composition in the DSO Cognitive Architecture

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES