Welcome to Image Discrimination Task!
Thank you for participating in our study. We are developing advanced computer vision systems designed to mimic human visual perception. To train these systems effectively, we need to understand how people solve visual problems especially in complex or ambiguous situations. Your participation helps us compare human and AI decision-making, and contributes to future advances in robotics, assistive technology, and visual cognition research.
In this experiment, you will complete a series of visual discrimination tasks. As you do, you will be working alongside other participants, including both humans and artificial agents. The artificial agents are trained on different conditions and thus will exhibit different capabilities across various image discrimination tasks. On each trial:
- Initial decision phase (5 seconds). You will see an image on the left and three candidate matches on the right. Your task is to choose which right-hand image best matches the one on the left. Once selected, your choice will be highlighted with a red border.
- Decision review phase (5 seconds). After the first phase, you will see the choices made by other participants including both humans and AI agents. You will then have the opportunity to revise your answer. If you change your response, your new selection will be highlighted with a green border. If you do not wish to change your response, click the same image again. Please note that you will not be made aware of the decision changes made by other participants at this stage.
Here are some other important things to note:
- Timing and compensation policy. This study is synchronized across participants, so answering quickly does not speed up the experiment. Likewise, the experiment will proceed even if you do not respond. To be eligible for compensation, you must respond in at least 80% of both decision phases across all trials. You are additionally forbidden from using the browser navigation controls (Back, Forward, Backspace key, etc). Doing so will automatically terminate the study with no compensation.
- Structure and duration. The task consists of 6 sets of 18 trials. Each trial is separated by a 1-second pause, and each set by a 30-second break.
- Post-task questionnaire. After completing the trials, you will answer a short questionnaire about your experience and perceptions. There is no time limit, but all questions must be completed to finalize your participation.
- Contact and withdrawal. If at any point you feel uncomfortable or wish to withdraw, you are free to do so by simply leaving this website. Please contact us at: maykristine.carlon[at]riken.jp for any concerns or questions.
Start the Image Discrimination Task Experiment