Experiments

Are People Confident When Artifical Neural Networks Are Confident?

State of the art artificial neural networks that are supposed to model the human visual system are disappointingly inconsistent with how humans classify images (Geirhos et al., 2020, Beyond accuracy: quantifying trial-by-trial behaviour of CNNs and humans by measuring error consistency). To further investigate this gap between artifical neural networks and humans, we want to see whether humans and neural networks are confident whenever they classify consistently and unconfident when they are inconsistent.

The experiment takes place at Sand 6, Room F118. There are no further exclusion criteria and left-handers are welcome (only criteria are: age 18-65, visual aids such as glasses or contact lenses if necessary). There will be 2 sessions each for 1 hours so that you can gain 2 Versuchspersonenstunden in total plus there will be additional monetary incentives during the experiment. Feel invited to join the experiment and please write an email to Kathi (katharina.deckenbach@student.uni-tuebingen.de), Ivo (ivo.adam@student.uni-tuebingen.de), or Sascha (sascha.meyen@uni-tuebingen.de)!

Accumulation of Early Visual Evidence

Visual processing is often understood as a drift diffusion model (Ratcliff, 1978, A Theory of Memory Retrieval). Until today, these models use constant drift rates suggesting that, over the time of looking at a visual stimulus, the rate at which our visual system gathers information is constant. We aim to update these models based on experimental data. We hypohtesize that, once our visual system picked up a bit of information about the stimulus, it adapts and gathers more information leading to an accelerated drift rate.

The experiment takes place at Sand 6, Room F118. There are no further exclusion criteria and left-handers are welcome (only criteria are: age 18-65, visual aids such as glasses or contact lenses if necessary). There will be 3 sessions each for 2 hours so that you can gain 6 Versuchspersonenstunden in total plus there will be additional monetary incentives during the experiment. Feel invited to join the experiment and please write an email to Madeleine (madeleine.soukup@student.uni-tuebingen.de) or Sascha (sascha.meyen@uni-tuebingen.de)!


COMPENSATION

You can collect money (10 € per hour) if you participate in behavioural experiments at our lab and (12 € per hour) for EEG and fMRI experiments. If you are a cognitive science student, you can also receive VPH (Versuchspersonenstunden). Psychology students can also receive VPH (Versuchspersonenstunden) for participating in experiments, but only those experiments for which the "Versuchsleiterin" is Kriti Bhatia.