Humans learning a complex task are picky and sticky

Tiago Quendera, Zachary F. Mainen, Champalimaud Foundation, Portugal; Dongrui Deng, Xi’an Jiaotong University, China; Mani Hamidi, University of Tubingen, Germany; Mattia Bergomi, Veos Digital, Italy; Gautam Agarwal, Claremont Colleges, United States

Posters 3 Poster

Cognitive science

Pacific Ballroom H-O

Presentation Time:
Sat, 27 Aug, 19:30 - 21:30 Pacific Time (UTC -8)

While neural networks approach human levels of perfor- mance in many complex tasks, they require much more training than humans. This may be because only humans can infer and apply generalizable principles from prior ex- perience (Lake, Ullman, Tenenbaum, & Gershman, 2017). However, the statistics that underlie the human learning process are poorly understood and hard to investigate in the large state spaces found in most complex tasks (van Opheusden & Ma, 2019). We thus designed a cog- nitive task whose potential solutions are few enough for subjects to densely sample policy space, but complex enough to compel intelligent search. We launched the game as a smartphone-based app (hexxed.io) to collect data from 10k human participants. We find that unlike re- inforcement learning agents (Deep-Q Networks ; DQNs), humans (1) search a highly restricted subset of the pol- icy space; (2) attempt even poor solutions many times before discarding them; (3) arrive at the optimal policy suddenly and unpredictably with a “leap of insight”. Our data suggest a “top-down” learning process by which hu- mans propose explanatory solutions which they replace only upon collecting sufficient evidence to the contrary, in contrast to the “bottom-up” learning of DQNs that as- sociates states with rewarding actions

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
2022 Conference on Cognitive Computational Neuroscience
No resources available.
Session P-3
P-3.1: The role of semantics in similarity judgements of scene stimuli
Katerina Marie Simkova, Jasper van den Bosch, University of Birmingham, United Kingdom; Ian Charest, Université de Montréal, Canada
P-3.2: Voxel-wise Encoding Models with Hierarchical Task-optimized Brain Atlas
Huzheng Yang, Shi Gu, University of Electronic Science and Technology of China, China; Yuanning Li, ShanghaiTech University, China
P-3.3: Hidden knobs: Representations for flexible goal-directed decision-making
Romy Froemer, Amitai Shenhav, Brown University, United States; Sebastian Gluth, University of Hamburg, Germany
P-3.4: AI-based modeling of brain and behavior: Combining neuroimaging, imitation learning and video games
Anirudha Kemtur, Francois Paugam, Basile Pinsard, Pravish sainath, Yann Harel, Maximilien Le clei, Julie Boyle, Karim Jerbi, Pierre Bellec, University of montreal, Canada
P-3.5: Intrinsic ionic dynamics, oscillations, and resonance are reflected in and can be extracted from neuronal spike-train cross-correlations
Rodrigo FO Pena, Horacio G Rotstein, New Jersey Institute of Technology and Rutgers University, United States; Martín V Ibarra, Universidad Nacional de la Patagonia San Juan Bosco & CONICET, Argentina
P-3.6: Prediction of brain regions from single channel ECoG signals by deep learning
Ryosuke Negi, Tsukuba Univesity, Japan; Masaru Kuwabara, Ryota Kanai, Araya, Inc, Japan
P-3.7: Image Embeddings Informed by Natural Language Significantly Improve Predictions and Understanding of Human Higher-level Visual Cortex
Aria Wang, Michael Tarr, Leila Wehbe, Carnegie Mellon University, United States
P-3.8: An Efficient Search for Novel Behavioral Strategies in a Vast Program Space
Tzuhsuan Ma, Ann Hermundstad, HHMI Janelia Research Campus, United States
P-3.9: The Individualized Neural Tuning Model: Precise and generalizable cartography of functional architecture in individual brains
Ma Feilong, Guo Jiahui, Yaroslav O. Halchenko, James V. Haxby, Dartmouth College, United States; Samuel A. Nastase, Princeton University, United States; M. Ida Gobbini, University of Bologna, Italy
P-3.10: Brain State Dynamics Underlying False Alarms
Bikash Sahoo, Adam Snyder, University of Rochester, United States
P-3.11: Responses in an orientation recall task are generated by taking expectations of distributional beliefs
Peter Vincent, Athena Akrami, Sainsbury Wellcome Centre, United Kingdom; Maneesh Sahani, Gatsby Computational Neuroscience Unit, United Kingdom
P-3.12: The best advice you can give.
Sevan Harootonian, Mark Ho, Nastasia Klevak, Yael Niv, Princeton University, United States
P-3.13: Modality specificity and generality in the hierarchical levels of cognitive control
Taehyun Yoo, Hyeon-Ae Jeon, Daegu Gyeongbuk Institute of Science and Technology (DGIST), Korea (South)
P-3.14: Sequential object-based attention for robust visual reasoning
Hossein Adeli, Seoyoung Ahn, Gregory Zelinsky, Stony Brook University, United States
P-3.15: Deep convolutional neural networks fail to classify images ‘in the wild’
Michelle Greene, Jennifer Hart, Bates College, United States
P-3.16: Modeling pain in the brain with conditional variational autoencoder
Sungwoo Lee, Jihoon Han, Choongwan Woo, Sungkyunkwan Univercity / Institute for Basic Science, Korea (South)
P-3.18: Using Massive Individual fMRI Movie Data to Align Artificial and Brain Representations in an Auditory Network
Maëlle Freteault, Université de Montréal, IMT Atlantique, Canada; Basile Pinsard, Julie Boyle, Centre de Recherche de l'Institut Universitaire de Gériatrie de Montréal, Canada; Pierre Bellec, Université de Montréal, Canada; Nicolas Farrugia, IMT Atlantique, France
P-3.19: Conversion of ConvNets to Spiking Neural Networks With Less Than One Spike per Neuron
Javier López-Randulfe, Nico Reeb, Alois Knoll, Technical University of Munich, Germany
P-3.20: Brain-to-Brain Linguistic Coupling in Natural Conversations
Zaid Zada, Samuel Nastase, Ariel Goldstein, Uri Hasson, Princeton University, United States
P-3.21: Optimal encoding of prior information in noisy working memory systems
Hua-Dong Xiong, The University of Arizona, United States; Xue-Xin Wei, The University of Texas at Austin, United States
P-3.22: Angular gyrus responses show joint statistical dependence with brain regions selective for different categories
Mengting Fang, University of Pennsylvania, United States; Aidas Aglinskas, Stefano Anzellotti, Boston College, United States; Yichen Li, Harvard University, United States
P-3.23: Different computational strategies for different reinforcement learning problems
Pieter Verbeke, Tom Verguts, Ghent University, Belgium
P-3.24: Investigating individual differences in structure learning
Avinash Vaidya, David Badre, Brown University, United States
P-3.26: Revealing dimensions underlying the organization of observed actions
Zuzanna Kabulska, Angelika Lingnau, University of Regensburg, Germany
P-3.27: Human-like capacity limitation in multi-system models of working memory
Yudi Xie, Christopher Cueva, Guangyu Robert Yang, Massachusetts Institute of Technology, United States; Yu Duan, Aohua Cheng, Tsinghua University, China; Pengcen Jiang, University of Science and Technology of China, China
P-3.28: Are there “affect detectors” in the human limbic system? A multivariate analysis of intracranial single cell recordings
Alexander Lawriw, Christopher Cox, Louisiana State University, United States
P-3.29: Do Convolutional Neural Networks Model Inferior Temporal Cortex Because of Perceptual or Semantic Features?
Anna Truzzi, Rhodri Cusack, Trinity College Dublin, Ireland
P-3.30: Fixation duration variability increases with mind wandering during scene viewing
Kevin O'Neill, Kristina Krasich, Felipe De Brigard, Duke University, United States; Samuel Murray, Providence College, United States; James Brockmole, University of Notre Dame, United States; Antje Nuthmann, Kiel University, Germany
P-3.31: Individual auto-regressive models for long-term prediction of BOLD fMRI signal
François Paugam, Guillaume Lajoie, Pierre Bellec, Université de Montréal, Canada; Basile Pinsard, Centre de Recherche de l'Institut Gériatrique de Montréal, Canada
P-3.32: Face Pareidolia Selectivity in Macaque Face-Cells Does Not Reflect Perceived Faceness
Saloni Sharma, Kasper Vinken, Margaret Livingstone, Harvard Medical School, United States
P-3.33: Do deep convolutional neural networks accurately model representations beyond the ventral stream?
Dawn Finzi, Daniel Yamins, Kalanit Grill-Spector, Stanford University, United States; Kendrick Kay, University of Minnesota, United States
P-3.34: Rationalizing Behavior during Virtual Reality Navigation
Yizhou Chen, Baylor college of medicine, United States; Paul Schrater, University of Minnesota, United States; Dora Angelaki, New York University, United States; Xaq Pitkow, Rice University, Baylor college of medicine, United States
P-3.35: Neural Mechanisms of Credit Assignment for Inferred Relationships in a Structured World
Phillip Witkowski, Seongmin Park, Erie Boorman, University of California, Daivs, United States
P-3.36: Developmental differences in social brain responses during movie viewing
Angira Shirahatti, Leyla Isik, Johns Hopkins University, United States
P-3.37: Transfer learning in a 3D-CNN is beneficial for small sample sizes in HCP task data
Philipp Seidel, Jens V. Schwarzbach, Regensburg University, Germany
P-3.38: Gaze-centered spatial representations in human hippocampus
Zitong Lu, Julie Golomb, The Ohio State University, United States; Anna Shafer-Skelton, University of Pennsylvania, United States
P-3.39: Large Scale Resting-State Network Connectivities Predict Verbal Suggestibility
Yeganeh Farahzadi, Zoltan Kekecs, Eötvös Loránd University, Hungary
P-3.41: Lateral Inhibition Facilitates Sequential Learning in a Hippocampus-Inspired Auto-Associator
Benjamin Midler, James McClelland, Stanford University, United States
P-3.42: Latent dimensionality scales with the performance of deep learning models of visual cortex
Eric Elmoznino, Michael Bonner, Johns Hopkins University, United States
P-3.43: Towards Precise and Robust Hippocampus Segmentation using Self-Supervised Contrastive Learning
Kassymzhomart Kunanbayev, Donggon Jang, Jeongwon Lee, Dae-Shik Kim, KAIST, Korea (South)
P-3.44: Effects of Predictability and Controllability on Pain Perception
Marie Habermann, Christian Büchel, University Medical Center Hamburg-Eppendorf, Germany
P-3.45: A mathematical framework for bridging Marr’s levels
Anja Meunier, Moritz Gosse-Wentrup, University of Vienna, Austria
P-3.46: Precision-weighted evidence integration predicts time-varying influence of memory on perceptual decisions
Maria Khoudary, Megan Peters, Aaron Bornstein, University of California, Irvine, United States
P-3.47: How Well Do Contrastive Learning Algorithms Model Human Real-time and Life-long Learning?
Chengxu Zhuang, Violet Xiang, Daniel Yamins, Stanford University, United States; Yoon Bai, James DiCarlo, MIT, United States; Xiaoxuan Jia, Allen Institute, United States
P-3.48: Different Brain Mechanisms of Time Estimation Depending on Situational Information
Jungtak Park, Hyeon-Ae Jeon, Daegu Gyeongbuk Institute of Science and Technology (DGIST), Korea (South)
P-3.49: Uncovering the Spatiotemporal Dynamics of Goal-driven Efficient Coding with a Brain-supervised Sparse coding Network
Bruce Hansen, Isabel Gephart, Victoria Gobo, Colgate University, United States; Michelle Greene, Bates College, United States; David Field, Cornell University, United States
P-3.50: A New Computational Framework for Estimating Spatio-temporal Population Receptive Fields in Human Visual Cortex
Insub Kim, Eline Kupers, Kalanit Grill-Spector, Stanford University, United States; Garikoitz Lerma-Usabiaga, Basque Center on Cognition, Brain and Language, Spain; Won Mok Shim, Sungkyunkwan University, Korea (South)
P-3.51: Economically expanding internal models in human density estimation
Tianyuan Teng, Hang Zhang, Peking University, China; Li Kevin Wenliang, University College London, United Kingdom
P-3.52: The geometry of cognitive maps under dynamic cognitive control
Seongmin Park, Maryam Zolfaghar, Jacob Russin, Douglas Miller, Randall O’Reilly, Erie Boorman, University of California, Davis, United States
P-3.53: Reconstruction of line illusion from human brain activity
Fan Cheng, Tomoyasu Horikawa, Advanced Telecommunications Research Institute International(ATR), Japan; Kei Majima, Yukiyasu Kamitani, Kyoto University, Japan
P-3.54: Statistical inference on representational geometries
Heiko Schütt, Alexander D. Kipnis, Nikolaus Kriegeskorte, Columbia University, United States; Jörn Diedrichsen, Western University, Canada
P-3.55: Network Architecture of Cortex and Cerebellum for Supporting Super-learning
Serhat Çağdaş, Yalova University, Turkey; Ismail Akturk, Ozyegin University, Turkey; N. Serap Şengör, İstanbul Technical University, Turkey
P-3.56: Humans learning a complex task are picky and sticky
Tiago Quendera, Zachary F. Mainen, Champalimaud Foundation, Portugal; Dongrui Deng, Xi’an Jiaotong University, China; Mani Hamidi, University of Tubingen, Germany; Mattia Bergomi, Veos Digital, Italy; Gautam Agarwal, Claremont Colleges, United States
P-3.57: Leaving alternatives behind: A theoretical and experimental investigation of the role of mutual inhibition in shaping choice
Xiamin Leng, Romy Frömer, Thomas Summe, Amitai Shenhav, Brown University, United States
P-3.58: Orthogonal neural encoding of targets and distractors supports cognitive control
Harrison Ritz, Amitai Shenhav, Brown University, United States
P-3.59: Identifying transfer learning in the reshaping of inductive biases
Anna Székely, Wigner Research Centre for Physics // Budapest University of Technology, Hungary; Balázs Török, Mozalearn Ltd., Hungary; Dávid Gergely Nagy, Gergő Orbán, Wigner Research Centre for Physics, Hungary; Mariann M. Kiss, Dezső Németh, Eötvös Lóránd University, Hungary; Karolina Janacsek, University of Greenwich, United Kingdom
P-3.60: The Neural Representation of Real-World Object Size in Natural Images
Andrew Luo, Leila Wehbe, Michael Tarr, Margaret Henderson, Carnegie Mellon University, United States
P-3.61: Component Activity States Underlying Memory Reactivation in the Posterior Medial Cortex
Yoonjung Lee, Hongmi Lee, Janice Chen, Johns Hopkins University, United States
P-3.62: Time cell encoding is decoupled from time perception in deep reinforcement learning agents
Ann Zixiang Huang, Dongyan Lin, Blake Richards, McGill University, Quebec AI Institute (Mila), Canada
P-3.63: Syntax in working memory using a simple plastic attractor
Lin Sun, Imperial College London, United Kingdom; Sanjay G. Manohar, University of Oxford, United Kingdom
P-3.64: Relating covariability in visual cortex to natural image statistics
Amirhossein Farzmahdi, Ruben Coen-Cagli, Albert Einstein College of Medicine, United States
P-3.65: Learning efficient attractor-based working memory representations in heterogeneous environments
Tahra L Eissa, Zachary P Kilpatrick, University of Colorado Boulder, United States
P-3.66: Attractor dynamics account for decision uncertainty in macaque prefrontal cortex
Siyu Wang, Rossella Falcone, Barry Richmond, Bruno Averbeck, National Institute of Mental Health, United States
P-3.67: A Characterization of the Neural Representation of Confidence during Probabilistic Learning
Tiffany Bounmy, NeuroSpin, CEA, INSERM, Université Paris-Saclay, Université de Paris, France; Evelyn Eger, Florent Meyniel, NeuroSpin, CEA, INSERM, Université Paris-Saclay, France
P-3.68: Representational dynamics of listened and imagined musical sound sequences
David Ricardo Quiroga-Martinez, Robert Knight, Helen Wills Neuroscience Institute and the Department of Psychology, UC Berkeley, United States; Leonardo Bonetti, Center for Eudaimonia and Human Flourishing, Linacre College & Department of Psychiatry, University of Oxford, United Kingdom; Peter Vuust, Center for Music in the Brain, Aarhus University and the Royal Academy of Music, Denmark, Denmark
P-3.69: Analysis of Transformer attention in EEG signal classification
Philipp Thölke, Karim Jerbi, University of Montreal, Canada
P-3.70: Superstitious learning of abstract order from random reinforcement
Yuhao Jin, Jacqueline Gottlieb, Vincent Ferrera, Columbia University, United States; Greg Jensen, Reed College, United States
P-3.71: Overcoming the Failure of Neoclassical Economics to Capture Excessive Demand: A Learning-to-Neuroforecast Experimental Approach
John Haracz, Indiana University, United States
P-3.72: Probing population codes and circuit dynamics of probabilistic learning
Nuttida Rungratsameetaweemana, The Salk Institute for Biological Studies, United States; Shruti Kumar, Javier Garcia, US Combat Capabilities Development Command Army Research Laboratory, United States
P-3.73: Occluded object completion occurs in full across human visual cortex but emerges gradually across layers of CORnet-S
David Coggan, Frank Tong, Vanderbilt, United States