1 of 7

Hackathon Project: Autis(CASP)

William Doan,

Ashraful Islam,

Mahd Malik

Autis(CASP)

Confidential

Copyright ©

1

2 of 7

Autis(CASP)

An application that uses logic and facts to determine whether or not a person is likely to be autistic and determines the severity level.

  • Based on a paper one of our members co-authored which was recently accepted into IEEE Transactions on Multimedia.
  • Supplemented with research from DSM-5.

2

Confidential

Copyright ©

3 of 7

The Problem

As the demand for mental healthcare balloons, long wait times and inaccessibility hinders the directives of the Hippocratic Oath.

Autis(CASP) solves this by using logic to efficiently and swiftly screen and determine the presence and extent of autism in a user.

3

3

Confidential

Copyright ©

4 of 7

The Application

Part 1: Gathering Info

Part 2: Sending Query

The chatbot first asks questions related to the 7 identified categories of autism, and stores the answers to send later.

The chatbot then sends this to python which sends a query to s(CASP).

Our application starts on the main page, where the user can send messages to our chatbot or have suggestions on which message to send.

4

Confidential

Copyright ©

5 of 7

The Team

Mahd Malik

William T. Doan

Ashraful Islam

Created base chatbot, handled queries from chatbot to python and back, integrated python with s(CASP).

Worked on the UI, logo, images and helped with integrating Google Gemini AI API and writing the AI prompt.

Worked on the main logic engine using s(CASP) providing his past, award-winning, research to assist with the queries.

Text Here�Title

Text Here�Title

5

Confidential

Copyright ©

6 of 7

What Makes

Us Unique

  • We are the first mental health screening app of its kind to incorporate autism diagnoses and levels of severity using logic and an LLM.

  • We based our work on the advancements made in [1]. Fortunately, a co-author on [1] is a member of our team.

[1] S. Deng et al., “Hear me, see me, understand me: Audio-visual autism behavior recognition,” *arXiv preprint arXiv:2406.02554v1*, Mar. 2024.

6

Confidential

Copyright ©

7 of 7

Important

Links

7

Confidential

Copyright ©