1/n Our work on a perceptual confirmation bias is finally out. Why do different perceptual decision-making studies find different temporal biases? What is their computational basis? Joint work with @wrongu, @ankani, @jake and Jeff Beck.

RT @FUTURUMCareers: In our latest collection of #free #education #resources, we explore #cognitive #science with Professor Ralf Haefner fro….

Blazing-fast image creation – using just your voice. Try Grok Imagine.

RT @pebenjamters: “How does the primate brain combine generative and discriminative computations in vision?”. The preprint of our @CogCompN….

RT @GrowingUpInSci: Next week (Nov 16), @SussilloDavid will tell his unofficial story. Among all that we have heard over the years, his is….

RT @neuromatch: We believe. 💎 Everyone should be able to read and publish research for free. 🌍 Research publishing should be commonly owned….

RT @TonyZador: Faculty search in COMPUTATIONAL/THEORETICAL NEURO at CSHL. Please spread the word!.

@sabyashiv1729 @DMWolpert @lengyel_m (Co-)first author refs only:.Approx causal inference w/ @maddox_lab.Differences in causal inference between ASD observers & controls w/ @JP__NOEL &Dora Angelaki.Active sensing w/ @AnkaniChattoraj. more soon.

Congratulations to @sabyashiv1729 for defending his PhD! 🥂🍾.Excellent and wide-ranging work on approximate inference in motion processing, cue integration and active sensing. and soon becoming a postdoc w/@DMWolpert & @lengyel_m

RT @sabyashiv1729: Great experience collaborating with @JP__NOEL, @AngelakiLab and @haefnerlab! .Modeling takeaways in thread.

RT @CogCompNeuro: There are only TEN DAYS left to submit your proposal for a GAC at CCN 2022!. “A GAC?“, you say. “What’s a GAC?”. Allow us….

Check out our #VSS2022 poster on the limits and time scales of perceptual confirmation bias during visual evidence integration on Sun AM with @xuan_wen_99 @AnkaniChattoraj. Based on with @wrongu @jcbyts and Jeff Beck.

Interested in how and why perceptual confirmation bias leads to overconfidence during evidence integration? Check out @AnkaniChattoraj’s poster on Sun AM #VSS2022 Continuation of with @MSnarskis.

RT @fhuszar: Postdoc on Probabilistic Machine Learning with @ZoubinGhahrama1 and me in Cambridge. If you know anyone interested, please enc….

RT @nencki: We are happy to announce Nencki School of Ideas in Neuroscience! We are going to discuss exciting theories, have a philosopher'….

@wrongu @ankani @jake 8/n Many thanks to @summerfield, @realrickborn, @nienborglab, @aplabur, @azylb, Duje Tadin and many others for helpful comments.

@wrongu @ankani @jake 7/n Our data are not explainable by existing alternative accounts: neither integration-to-bound nor attractor-dynamics within a decision area are sensitive to the balance of sensory and category information - only to their combination.

@wrongu @ankani @jake 6/n We verified a key prediction of our model: the bias of the same observer should change as the balance of sensory and category information is changed - keeping everything else constant - using an orientation discrimination task.

@wrongu @ankani @jake 5/n High category information means that early evidence is highly predictive of later evidence within a trial. When sensory information is low, perception is biased towards earlier beliefs. Hence, confirmation bias.

@wrongu @ankani @jake 4/n Animated example:

@wrongu @ankani @jake 3/n Example: Left: low sensory information (hard to detect orientation on each frame) and high category information (all frames have the same orientation). Right: high sensory (orientation clear on each frame) and low category information (orientations flip back and forth)

@wrongu @ankani @jake 2/n Why do observers sometimes have a primacy bias? And sometimes weigh evidence equally or have a recency bias? Differences in species? Training? Stimulus modality? We propose the key factor is the balance between “sensory” and “category” information.