Netherlands

The creation of sophisticated AI systems that are able to process and produce images and text creates new challenges in assessing the capabilities of those systems. We adapt a behavioral paradigm from developmental psychology to characterize the counting ability of a model that generates images from text. We show that three model scales of the Parti model (350m, 3B, and 20B parameters respectively) each have some counting ability, with a significant jump in performance between the 350m and 3B model scales. We also demonstrate that it is possible to interfere with these models&#39; counting ability simply by incorporating unusual descriptive adjectives for the objects being counted into the text prompt. We analyze our results in the context of the knower-level theory of child number learning. Our results show that we can gain experimental intuition for how to probe model behavior by drawing from a rich literature of behavioral experiments on humans, and, perhaps most importantly, by adapting human developmental benchmarking paradigms to AI models, we can characterize and understand their behavior with respect to our own.

**Authors:**

Sunayana Rane: Princeton University; Alexander Ku: Google DeepMind; Jason Baldridge: Google; Ian Tenney: Google Research; Tom Griffiths: Princeton University ; Been Kim: Google

CogSci 2024

Can Generative Multimodal Models Count to Ten?

It is my great pleasure and honor to welcome you to the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), held in San Diego, California, USA, July 2–7, 2026. ACL 2026 continues our fields tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.

It is exciting to see the many developments in the field reflected at our conference, such as the advancement of large language models (born out of our community) into many application directions. This is not without challenges, partly due to unsolved problems, and partly due to misuse of the technology. All this is reflected well at ACL 2026. The conference also reflects that the field is not only about advancing technology, but also the science of fundamentally understanding language and how it is use by many different speaker communities. <br>
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) <br><br><br>
<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</button></html>
<br><br><br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>Access credentials will be sent to your email from Underline -  subject line "Welcome to ACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.aclweb.org/registration/) first.

Registration Required

ACL 2026

The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) will take place in San Diego, California, from July 2 - 7, 2026.

The creation of sophisticated AI systems that are able to process and produce images and text creates new challenges in assessing the capabilities of those systems. We adapt a behavioral paradigm from developmental psychology to characterize the counting ability of a model that generates images from text. We show that three model scales of the Parti model (350m, 3B, and 20B parameters respectively) each have some counting ability, with a significant jump in performance between the 350m and 3B model scales. We also demonstrate that it is possible to interfere with these models' counting ability simply by incorporating unusual descriptive adjectives for the objects being counted into the text prompt. We analyze our results in the context of the knower-level theory of child number learning. Our results show that we can gain experimental intuition for how to probe model behavior by drawing from a rich literature of behavioral experiments on humans, and, perhaps most importantly, by adapting human developmental benchmarking paradigms to AI models, we can characterize and understand their behavior with respect to our own.

**Authors:**

Sunayana Rane: Princeton University; Alexander Ku: Google DeepMind; Jason Baldridge: Google; Ian Tenney: Google Research; Tom Griffiths: Princeton University ; Been Kim: Google

technical paper

The 46th Annual Meeting of the Cognitive Science Society was an in-person meeting held in Rotterdam, The Netherlands at the Postillion Hotel & Conference Centre.

**ON-DEMAND PROGRAM ACCESS AFTER THE CONFERENCE**<br>
Recordings of the invited program are now available. You can access them by clicking on the 'Schedule' icon on the left. Select view by week and navigate to the day and time slot of the recording you wish to view. Click on the time slot to access the recording.

Recordings available (click on each recording to view):

**Thursday, July 25**<br>
0900: [Keynote Speaker Morgan Barense on Enhancing real-world event memory](https://underline.io/events/465/sessions/17997/lecture/99084-dynamics-between-minds-and-the-environment)<br>
1000: [Gleitman Award Winner Isabelle Dautriche on Language Foundations: Insights from acquisition, communication, cognition, and more](https://underline.io/events/465/sessions/18000/lecture/99085-gleitman-talk)<br>
1415: [Invited Symposium: Dynamics between minds and the environment](https://underline.io/events/465/sessions?eventSessionId=18016&searchGroup=lecture)<br>
1700: [Rumelhart Prize Presentation Speaker Alison Gopnik on Exploit, explore, empower: Three ages and three intelligences](https://underline.io/events/465/sessions/18028/lecture/99161-rp1-rumelhart-prize-presentation)<br>

**Friday, July 26**<br>
0900: [C.L. de Carvalho-Heineken Prize Keynote Speaker Kia Nobre on Focusing in memory](https://underline.io/events/465/sessions/18035/lecture/99162-hp1-c-l-de-carvalho-heineken-prize-keynote-address)<br>
1030: [Rumelhart Symposium: Childhood as exploration](https://underline.io/events/465/sessions/18038/lecture/100346-childhood-as-exploration)<br>
1415: [Elman Prize Symposium](https://underline.io/events/465/sessions?eventSessionId=18053&searchGroup=lecture)<br>
1600: [Invited Symposium: Dynamics between minds](https://underline.io/events/465/sessions?eventSessionId=18065&searchGroup=lecture)<br>
1745: [Keynote Speaker Andrea E. Martin on Neural dynamics encode the structure and statistics of language](https://underline.io/events/465/sessions/18077/lecture/99271-neural-dynamics-encode-the-structure-and-statistics-of-language)<br>

**Saturday, July 27**<br>
0900: [Keynote Speaker Gregor Schöner on How higher cognition emerges from the dynamics of strongly interacting neural populations](https://underline.io/events/465/sessions/18083/lecture/99273-dynamics-within-the-mind)<br>
1000: [Glushko Talks](https://underline.io/events/465/sessions?eventSessionId=18086&searchGroup=lecture)<br>
1500: [Invited Symposium: Dynamics within minds](https://underline.io/events/465/sessions?eventSessionId=18100&searchGroup=lecture)<br>


#### **CogSci 2024 Program Book**
<div style="position:relative;padding-top:0;width:900px;height:500px;"><iframe style="position:absolute;border:none;width:100%;height:100%;left:0;top:0;" src="https://online.fliphtml5.com/ebtyf/cble/"  seamless="seamless" scrolling="no" frameborder="0" allowtransparency="true" allowfullscreen="true" ></iframe></div>


**The Program Book can be downloaded [here](https://drive.google.com/file/d/1yRVp1cBqRnAbdYi4GXWV_pZeyWNsRquv/view?usp=sharing).**

#### **Letter from the President**

Welcome to the 2024 Meeting of the Cognitive Science Society in Rotterdam! CogSci 2024 brings together a large community of cognitive scientists who have traveled here from around the world, as well as a smaller group of remote presenters. I would like to extend an especially warm welcome to our first-time CogSci meeting attendees, and hope that 
their experience inspires them to join our future meetings for many years to come.

I want to thank this year’s conference co-chairs, Larissa K Samuelson, Stefan Frank, Mariya Toneva, Allyson Mackey, and Eliot Hazeltine, for putting together an amazing series of invited talks and symposia around the theme of Dynamics of Cognition. The co-chairs also deserve credit for handling a record-breaking number of submissions and producing a truly exciting conference program 

Beyond the invited talks and symposia dedicated to this year’s conference theme, I would like to highlight the presentations and events honoring the 
Rumelhart, Elman, Gleitman and C .L . de Carvalho Heineken prize winners, including the Rumelhart reception on Thursday. As you may know, we are recording all of the invited talks and symposia, and all prize-winners’ talks, and will be making this content available after the conference. 

The Cognitive Science Society relies on the volunteer efforts of our Governing Board members who work on matters related to our membership; conference policies; diversity and inclusion, international, and outreach initiatives; prizes and much more. We welcome interest from Society members who would like to become more involved in what we do. The 
Society is deeply grateful to Rick Dale and Andrea Bender, who serve as editors of the two Society journals, Cognitive Science and Topics in Cognitive Science respectively. We especially acknowledge our editors’ generosity in donating their stipends each year to D&I and outreach efforts. 

This year’s Annual Meeting depends critically on the support of Marischal De Armond and his team at Podium Conference Specialists: Sharon Zwack, Cendrine De Vis, Sarah-Kate Burke, and Rachel Waller worked tirelessly to secure the venue and handled a multitude of logistical and organizational issues, from arranging coffee breaks and receptions to making sure that a huge number of posters fits comfortably into the available space. Finally, and importantly, I would like to acknowledge the Cognitive Science Society Executive Officer, Erica Wojcik, for managing the complex and ever expanding sphere of our Society’s activities.

I hope you enjoy this year’s meeting and the many different opportunities to engage with the vibrant community of cognitive scientists gathered here. I also hope you get a chance to explore the great city of Rotterdam with its many attractions, museums, restaurants and scenic views. If you’d like to talk more about our Society, or simply want to say hello, please stop by!

![](https://assets.underline.io/markdown_image/1/image/3c66d0e21fff30e204df1138ffd15173.jpeg)

**Anna Papafragou**  
President, Cognitive Science 
Society 2023-2024

**CODE OF CONDUCT**<br>
By attending the CogSci 2024 Conference, you are required to adhere to the society’s [Code of Conduct](https://drive.google.com/file/d/16-6KkptF0Gn3ZYGJDlpqpwTdImw45Ng0/view?usp=drive_link).

**ABOUT THE COGNITIVE SCIENCE SOCIETY**<br>
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.
<br><br>

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

It looks like you are not registered for this event. 

To access the site please register [**here**](https://cognitivesciencesociety.org/registration/). 

Please register!

The 46th Annual Meeting of the Cognitive Science Society presents the latest research across cognitive science and highlights the theme of Cognition in Context.

Recent advancements in Artificial Intelligence (AI) have arguably enhanced human performance instead of supplanting it. Here we analyse 2.8 million decisions by elite chess players, a field emblematic of AI's application due to its complexity and objective measurability. We identify two AI milestones that correspond with substantial enhancements in top chess players' performance quality over the past four decades: the introduction of personal computers (PCs) and internet access in the late 1990s, and the advent of deep neural networks for chess in the late 2010s. The impact of these technologies, however, varied by age group; adult elite players derived considerable benefits from neural network-based chess computers, whereas younger top players were more influenced by the widespread availability of knowledge and PCs. Our findings underscore AI's potential to amplify human proficiency in complex tasks, highlighting the importance of tailored technological integration among elite performers.0

Advancing the (Elite) Grandmasters: AI's Role in Enhancing Chess Expertise

This study investigates how people assign blame to autonomous vehicles (AVs) when involved in an accident. Our experiment (N = 2647) revealed that people placed more blame on AVs than on human drivers when accident details were unspecified. To examine whether people assess major classes of blame-relevant information differently for AVs and humans, we developed a causal model and introduced a novel concept of prevention effort, which emerged as a crucial factor for blame judgement alongside intentionality. Finally, we addressed the “many hands” problem by exploring how people assign blame to entities associated with AVs and human drivers, such as the car company or an accident victim. Our findings showed that people assigned high blame to these entities in scenarios involving AVs, but not with human drivers. This necessitates adapting a model of blame for AVs to include other agents and thus allow for blame allocation “outside” of autonomous vehicles.0

Are autonomous vehicles blamed differently?

Children learn powerful internal models of the world around them from a few years of egocentric visual experience. Can such internal models be learned from a child's visual experience with highly generic learning algorithms or do they require strong inductive biases? Recent advances in collecting large-scale, longitudinal, developmentally realistic video datasets and generic self-supervised learning (SSL) algorithms are allowing us to begin to tackle this nature vs. nurture question. However, existing work typically focuses on image-based SSL algorithms and visual capabilities that can be learned from static images (e.g. object recognition), thus ignoring temporal aspects of the world. To close this gap, here we train self-supervised video models on longitudinal, egocentric headcam recordings collected from a child over a two year period in their early development (6-31 months). The resulting models are highly effective at facilitating the learning of action concepts from a small number of labeled examples; they have favorable data size scaling properties; and they display emergent video interpolation capabilities. Video models also learn more robust object representations than image-based models trained with the exact same data. These results suggest that important temporal aspects of a child's internal model of the world may be learnable from their visual experience using highly generic learning algorithms and without strong inductive biases.0

Self-supervised learning of video representations from a child's perspective

Independent pieces of corroborating evidence should provide stronger support to a hypothesis than dependent pieces of evidence. Overlooking the inferiority of dependent relative to independent items of evidence can lead to a chain reaction of double-counting evidence, over-estimating the probability that the fact under consideration is true, and making wrongful decisions. Within fictitious scenarios, we investigate people’s sensitivity to the independency advantage. We assess their ability to integrate multiple items of evidence that come from (in)dependent sources who differ in reliability. We find that participants properly perceive dependencies when explicitly asked but fail to distinguish the probative value of dependent versus independent evidence in their belief updating. Still, individuals who perceive a strong dependence between sources treat the evidence as being more redundant. We find no dependency-related effects on participants’ individual Bayesian network model predictions. Potential reasons why participants perceive (in)dependencies and yet (mostly) fail to discount for them are discussed.0

Reasoning about (In)Dependent Evidence: A Mismatch between Perceiving and Incorporating Dependencies?

Contemporary formal models aim to capture group polarization as the result of deliberation between rational agents. Paradigmatic models do, however, rely on rather limited agents, casting doubt on the conclusion that group polarization can be rationally reconstructed. In this paper, we use a recently developed Bayesian agent-based model of deliberation to investigate this conclusion. This model avoids problems we identify in a group of influential Bayesian polarization models. Our case study shows that a simple mechanism produces realistic patterns of group polarization: limited exchange of evidence across a sparse social network.  We reflect on what our results mean for our formal understanding of rational group polarization.0

Rational Polarization: Sharing Only One’s Best Evidence Can Lead to Group Polarization

Humans seem to arbitrate between automatic and controlled processing by optimizing a trade-off between cognitive effort and performance. Previous research has described ways of how these costs and benefits can be quantified and how the trade-off between them can be performed. However, it remains unclear how the costs should be weighed relative to the benefits and how the cost of the arbitration mechanism itself factors in. Here, we derive measures for these separate factors from a single objective: the variational free energy. We demonstrate that by minimizing this objective, the trade-off between automatic and controlled processing as well as meta-control is optimized implicitly. As a proof of concept, we show that the congruency and proportion congruency effects in the Stroop task directly result from this optimization, given an environment with specific statistical regularities.0

A Rational Trade-Off Between the Costs and Benefits of Automatic and Controlled Processing

According to a common view in epistemology, a set of propositions is justified if it is coherent. Similarly, a new proposition should be accepted if it is coherent with the accepted body of beliefs. But what is coherence? And what, in turn, justifies the above claims? To answer these questions, various Bayesian measures of epistemic coherence have been proposed. Most of these measures are based on the prior probability distribution over the corresponding propositional variables. We criticize this ``static'' conceptualization of coherence and propose instead that the coherence of an information set is related to how well the information set responds when each of the propositions it contains is confirmed by an independent and partially reliable information source. The elaboration of this idea will show that the proposed ``dynamic'' perspective has several advantages and solves some open problems of coherentist epistemology. It also has implications for our understanding of reasoning and argumentation in science and beyond.0

A New Posterior Probability-Based Measure of Coherence

Metaphors are often used to intuitively communicate about movement. Here, expert pianists played two melodies while keeping eight different metaphors in mind, contrasting arousal level, valence direction, and metaphor type (action-related and emotion-related metaphors). Measures of keystroke timing and velocity were analyzed to assess the relative contribution of metaphor content and melodic note sequence to motor performance, alongside ratings of semantic similarity between metaphors. Using Bayesian multilevel models, results indicate that the arousal level of the metaphor has the most influence on keystroke force, average tempo, and tempo variability. Additionally, interactions with valence are seen for the timing measures, and for both valence and type in force. No effects of the melody sequence were found. Similarity ratings of metaphor pairs indicate that mental similarities largely mirror performance similarities. These findings show the potential effects of mental imagery on motor performance and have implications for teaching complex movements in practical settings.0

Metaphors in music performance: from semantics and motor performance to expressive communication

Emotions, including stress and anxiety, strongly influence cognition and learning experiences. This study investigates the impacts of stress on cognitive load during learning, considering baseline anxiety levels and fluctuating stress. With a focus on technology-based learning, a web-based HTML introduction module was used. Using a social stress test, 15 participants underwent a stressful situation during learning, while the control group of 15 were in a neutral condition. Results indicate significantly elevated stress levels in the experimental group throughout the experiment, with a corresponding decrease in learning performance. For high perceived difficulty, the stressed condition demonstrated a significant increase in response time compared to the control condition. In contrast, when experiencing low perceived difficulty, a significant difference in response time across conditions was not found. Findings emphasise the importance of managing stress in educational contexts to optimise learning outcomes in the evolving landscape of technology-based learning.0

The Effects of Stress and Anxiety in Technology-Based Learning Environments

Since their inception, drift diffusion models have been applied across a wide range of disciplines within psychology to uncover the mental processes that underlie perception, attention, and cognitive control. Our studies contribute to ongoing efforts to extend these models to abstract, social reasoning processes like moral or legal judgment. We presented participants with a set of social rules, while manipulating whether various
behaviors violated the rule’s letter and/or its purpose–––two independent standards by which to decide what constitutes a transgression. In this framework, cases that violate or comply with both a rule’s text and its purpose can be seen as congruent or ‘easy’ cases, and cases that elicit opposing verdicts as incongruent or ‘hard’ cases–––in a manner analogous to widely-studied conflict tasks in cognitive psychology. We recorded 34,573 decisions made by 364 participants under soft time pressure, and investigated whether hierarchical drift diffusion modeling could explain various behavioral patterns in our data. This approach yielded three key insights: (1) judgments of conviction were faster than judgments of acquittal owing to an overall bias (z parameter) toward conviction; (2) incongruent cases produced longer reaction times than congruent cases (an interference effect), due to differences in the rate of evidence accumulation (v parameter) across case-types; and (3) increases in the ratio of congruent-to-incongruent cases amplified the interference effect on reaction times, by fostering greater response caution—revealed by a larger threshold (or a parameter). Thus, our studies document dissociable effects of the drift diffusion components on rule-based decision-making, and illustrate how the cognitive processes that subserve abstract and social decision-making tasks, such as the enforcement of communal and legal rules, may be illuminated through the drift diffusion framework.0

Premium content

Downloads

Next from CogSci 2024

Advancing the (Elite) Grandmasters: AI's Role in Enhancing Chess Expertise

Similar lecture

Mechanistic Interpretability of Large-Scale Counting in LLMs through a System-2 Strategy