United States

Lake 2015 presented the Omniglot dataset to study how people generate and classify characters. They proposed a model for one-shot-learning of new concepts based on inferring compositionally structured generative models, and could transfer from familiar concepts to new ones. Their Bayesian-Programming-Language model was able to both classify and generate characters similar to adults. However, adults have years of experience, would a similar model apply to children without this prior knowledge? We introduce a new dataset called Omniglot_Jr., composed of Omniglot letters generated by children aged 3-6. Our results of training BPL with children&#39;s data in classification and generation find that BPL achieves higher classification accuracy and generates more adult looking letters. We propose the challenge of reproducing children&#39;s distinctive pattern of mistakes. Skills such as character recognition depend on child-like learning; this challenge should help us understand how that learning is possible, and how to simulate it in computational systems.

CogSci 2021

The Omniglot Jr. challenge; Can a model achieve child-level character generation and classification?

### 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.

Lake 2015 presented the Omniglot dataset to study how people generate and classify characters. They proposed a model for one-shot-learning of new concepts based on inferring compositionally structured generative models, and could transfer from familiar concepts to new ones. Their Bayesian-Programming-Language model was able to both classify and generate characters similar to adults. However, adults have years of experience, would a similar model apply to children without this prior knowledge? We introduce a new dataset called Omniglot_Jr., composed of Omniglot letters generated by children aged 3-6. Our results of training BPL with children's data in classification and generation find that BPL achieves higher classification accuracy and generates more adult looking letters. We propose the challenge of reproducing children's distinctive pattern of mistakes. Skills such as character recognition depend on child-like learning; this challenge should help us understand how that learning is possible, and how to simulate it in computational systems.

poster

Infants experience the world through their actions with objects and their interactions with other people, especially their parents. Prior research has shown that school-age children with hearing loss experience poorer quality interactions with typically hearing parents, and difficulties in controlling their visual attention. In the current study, we used mobile eye-tracking to investigate parent-child interactions in toddlers with and without hearing loss. Parents and toddlers engaged in a goal-directed, interactive task that involved inserting coins into a slot and required joint coordination between the parent and the child. We examined the visual behaviors of the toddlers and the scaffolding behaviors of the parents. In contrast to previous work, preliminary findings reveal a pattern of potential similarities between deaf and hearing toddlers or their parents.

Joint Action in Deaf and Hearing Toddlers: A Mobile Eye-Tracking Study

Cognitive control—the ability to execute goal-relevant responses in the presence of competing goal-irrelevant response alternatives—predicts academic achievement, delinquency, and occupational success. Assessing children's cognitive control is challenging due to the tedious nature of cognitive assessments and children’s low attention spans. This study examined whether a cognitive game-based assessment (GBA) may alleviate these challenges by investigating the construct and criterion validity of implementing game-based features into a traditional cognitive assessment, and the associations between GBA performance, academic achievement outcomes, and associated neural substrates in children ages 3-5. Performance on the GBA was significantly associated with performance on a traditional measure of cognitive control, functional brain connectivity, and mathematical and verbal test outcomes. Children also showed a stronger preference and higher ratings of enjoyment for the GBA compared to the traditional cognitive control assessment.

The Construct and Criterion Validity of a Cognitive Game-based Assessment: Cognitive Control, Academic Achievement, and Prefrontal Cortex Connectivity

Productive failure (PF) is a learning paradigm that reverses the standard order of instruction by asking students to solve problems prior to instruction. This paradigm has been shown to be effective for fostering student learning. To date, however, the role of student emotion in productive failure has not been investigated. In other paradigms, there is some evidence that failure elicits negative emotions and that these emotions can interfere with learning. This leads to a conundrum given productive failure’s positive effect on learning. To shed light on this, we report on results from a study (N = 48) in the productive failure paradigm. For the analysis, we used a mixed-methods approach to investigate the distribution of emotions in productive failure, how these changed across different instructional activities, and the relation between emotions and posttest performance.

Productive Failure and Student Emotions

Human-AI collaboration is an increasingly commonplace part of decision-making in real world applications. However, how humans behave when collaborating with AI is not well understood. We develop metacognitive bandits, a computational model of a human's advice-seeking behavior when working with an AI. The model describes a person's metacognitive process of deciding when to rely on their own judgment and when to solicit the advice of the AI. It also accounts for the difficulty of each trial in making the decision to solicit advice. We illustrate that the metacognitive bandit makes decisions similar to humans in a behavioral experiment. We also demonstrate that algorithm aversion, a widely reported bias, can be explained as the result of a quasi-optimal sequential decision-making process. Our model does not need to assume any prior biases towards AI to produce this behavior.

Explaining Algorithm Aversion with Metacognitive Bandits

Voice atypicalities are a characteristic feature of schizophrenia, often associated with core negative symptoms. A recent meta-analysis identified atypicalities in pitch, speech rate, and pauses. However, heterogeneity across studies was large and replications almost nonexistent. Further, it is not clear whether vocal patterns are directly related to the mechanisms underlying the disorder and could therefore be found across languages, or not.
In this study we implemented a more rigorously cumulative scientific approach by collecting and analyzing a large cross-linguistic corpus of voice recordings. We critically employed meta-analytic priors to systematically assess the replicability of previous findings, and modeled between-participants variability and cross-linguistic differences.
We replicate previous meta-analytic findings across all languages for reduced pitch variability, while increased pause duration and lower speech rate results were replicated only in some languages. Most atypical voice patterns, thus, seem not to be distinctive of schizophrenia in general, but more specifically situated in linguistic/cultural differences.

Vocal patterns in schizophrenia: toward a cumulative approach

Judging if a conclusion follows logically from a given set of
premises can depend much more on the believability than on
the logical validity of the conclusion. This so-called belief bias
effect has been replicated repeatedly for many decades now.
An interesting observation is, however, that process models for
deductive reasoning and models for the belief bias have not much
of an overlap - they have largely been developed independently.
Models for the belief bias often just implement first order logic
for the reasoning part, thereby neglecting a whole research field.
This paper aims to change that by presenting a first attempt at
substituting the first order logic components of two models for belief,
selective scrutiny and misinterpreted necessity, with two state
of the art approaches for modeling human syllogistic reasoning,
mReasoner and PHM. In addition, we propose an approach for
extending the traditionally dichotomous predictions to numerical
rating scales thereby enabling more detailed analysis. Evaluating
the models on a dataset published with a recent meta-analysis on
the belief bias effect, we demonstrate the general success of the
augmented models and discuss the implication of our extensions
in terms of the limitations of the current focus of research as well
as the potential for future investigation of human reasoning.

Unifying Models for Belief and Syllogistic Reasoning

With an increased reliance on technology to exchange social communication, social media platforms have become vehicles for humorous content — often taking the form of short internet videos (Vine or TikTok). Dynel (2016) hypothesizes that short internet videos may be effective because they capitalize on contrast to promote humor — i.e., consistent with incongruity theory (see Morreal, 2012). Specifically, incongruity theory assumes that humor arises from a contrast between an expectation and a perceived deviation from that expectation. To test the effects of incongruity in short internet videos, we manipulated linguistic and paralinguistic content in common short internet videos from the defunct social media platform - Vine. Results indicated that incongruence is extremely important for humor perception of short internet videos, as more linguistic and paralinguistic incongruity promoted higher social judgments of humor. Findings support that incongruity is an extremely effective tool for short internet videos to communicate humor.

Humorous Judgments of Incongruity in Short Internet Videos

In the development and analysis of interventions designed to improve student learning, it is important to consider potential influences of student behavior. Unproductive behaviors, such as “gaming the system,” have been studied for their potential impacts on the measurement and assessment of student knowledge within these interventions. Conversely, less attention has been given to factors that may influence gaming behavior. Gaming may be attributed to student-level traits, but could also be a temporary state brought on by systemic causes such as the perceived difficulty or presentation of a problem. We leverage prior research in this area to develop a data-driven measure of gaming and address whether this behavior occurs as a “state” or “trait” within a randomized controlled trial. We find that these factors are differential across the two study conditions, suggesting that context contributes to the underlying causes of gaming behavior.

State vs. Trait: Examining Gaming the System in the Context of Math Perception Tasks

People may engage in a joint activity (JA) to accrue material rewards or to help others. From a third-party perspective, the occurrence of JAs is thus ambiguous about the goals of the participating agents. We argue that the payoff structure of a JA (how costs and rewards are distributed) may help disambiguate these goals. Specifically, we hypothesize that an agent’s participation in a JA should be interpreted as prosocially motivated when its costs cannot be recouped by material rewards (disadvantageous payoff). We tested this hypothesis across three looking-time experiments with 12-month-olds. As predicted, infants expected a disadvantaged agent to behave altruistically towards her JA partner (Exp. 1). However, this expectation might be explained by a sensitivity to changes in overall reward distribution from familiarization to test (Exps. 2 & 3). Our results call for a re-evaluation of the role that payoff information plays in early goal attribution within JA contexts.

Do infants infer prosocial goals from disadvantageous payoffs in joint action?

Comparison settings (i.e. several stimuli introduced simultaneously) favor novel word learning and generalization. This study investigates the temporal dynamics of 6-year-olds solving strategies in a relational noun (e.g. “x is the dax for y”) comparison and generalization task with eye tracking data. We manipulated conceptual distance between the task’s items and recorded children’s performances and eye tracking data. We analyze and interpret solving strategies following the predictions made by two hypotheses, the Projection-First and Alignment-first. Eye tracking data clearly revealed that children, first, extract the relation from comparisons of items within a pair and search for a match for the extracted relation, which confirms the predictions of the projection-first hypothesis. Further analyses on error and correct trials suggest that errors occurred in the late, choice, phase of a trial.

Downloads

Next from CogSci 2021

Joint Action in Deaf and Hearing Toddlers: A Mobile Eye-Tracking Study

Similar lecture

Child-directed Listening: How Caregiver Inference Enables Children's Early Verbal Communication