Welcome Researchers to the PRAID Collaboration Portal.
On this Page:
Essential Resources
Access our shared documents and task planner with Microsoft tools.
- OneDrive – Access and share our project documents and files securely.
- Task Planner – Collaboratively plan and manage tasks.
White Paper Section Assignments
The problems are numbered but this is not a ranking; we can order things differently in the final version.
Section Format:
Summary of the Problem: description and explanation of why it is a problem
Current Approach(es): summary of current approach to the problem and why it fails; include description of one or two current examples and comprehensive reference list
Discussion: specific challenges; considerations that need to be addressed and those that can be tabled for now
Promising New/Alternative Approach: description of and justification (i.e. why this approach would address the problem) for new approach; include description of one or two examples of someone doing something close to this already; include relevant reference list
| Section | Problem/Question | Notes | Assignees |
| Intro | We don’t have/we need an operational or a minimal definition of AI perception (that takes the goals of the AI into account) | Minimal Condition: sensory uptake by a receptor meets a threshold that gets sent downstream | Cristina, Nancy |
| 1 | What are the impacts that the differences between human and AI perception will have on human/AI interactions? | Approach: If it’s similar, then that will help us understand AI perceptual judgements and their judgements will be aligned with human judgements? Challenge: Under/over estimating AI perception; assessing system safety in public sphere | Katharina; Cristina; Monica; Michael; Matthew |
| 2 | How can AI perception systems become sensitive to objects/situations not grounded in basic sensory features? | This is a narrow, perception-focused, statement of the more general problem of flexible adaptivity, which is a problem for AI in general. Approach: evolutionary/enactivist AI Challenge: significance-for, frame problem Approaches to Consider/Evaluate: Systems that can look at one and the same object and see that it is relevant to task A but not to task B. | Nancy; Ting; Sidney |
| 3 | We need a more scientific approach to developing and understanding perception systems. | Approach: Scientific process: attempt falsification! Inter-disciplinary Challenge: How do you assign/assess value to different AI perception systems? | Jeff; Monica |
White Paper Task Deadlines
MAX Words: 20,000 total, 6000 each section, 2000 introduction
| Date | Task |
| June 17-20 | Entire group (zoom) meeting to finalize section groups and assign writing tasks |
| June 24-July 5 | Group Meetings (groups organize their own meeting times) |
| Aug 19-30 | Group Meetings |
| Sept 16-20 | Everyone Meeting |
| Sept 21-30 | Group Meetings |
| October 31 | Deadline for first drafts of each section |
| Nov 4-8 | Everyone Meeting |
| November 2024 to April 2025 | Sections distributed amongst groups for review, edit, and integration |
| May 1 2025 | Whitepaper distributed to externals (who we’ve already identified) for review |
| July 1 2025 | Deadline for external reviews |
| July 1-30 2025 | Groups revise respective sections according to reviews |
| August 1 2025 | Final copy to group for final copyedit/review |
| August 31 2025 | Disseminate whitepaper |
All Finished?