Open Category Submission

356 – AI-Powered Language and Cognitive Processing Program for Individuals with Autism Spectrum Disorder

This project introduces an AI-powered educational application designed to enhance language and cognitive processing skills in individuals with Autism Spectrum Disorder (ASD). By leveraging Natural Language Processing (NLP), Automatic Speech Recognition (ASR), and Adaptive Learning AI, the solution offers a personalized and scalable learning experience that addresses the unique needs of individuals with ASD. It stimulates language use in non-verbal individuals with autism, improves echolalia (repetitive speech) in verbal individuals with autism, enhances linguistic structures for people with autism, and increases verbal fluency and language comprehension for social communication and interaction scenarios.

Key Innovations:
1. AI-Driven Personalisation – The system adapts dynamically to each user’s linguistic and cognitive abilities, providing tailored exercises and feedback.
2. Multimodal Learning Approach – Integrates speech processing, text analysis, and interactive visual cues to create a multi-sensory educational tool.
3. Real-Time Adaptive Feedback – Uses speech synthesis and interactive dialogues to provide instant feedback, reinforcing learning in a natural way.
4. Gamification and Engagement Features – Designed with interactive elements that motivate users with ASD to practice language in a fun, low-pressure environment.
5. Scalability and Accessibility – Deployable on mobile, web, and tablet platforms, with multilingual support to reach a broader user base.

Technical Principles:
- Uses deep learning models (transformers like BERT/GPT) to analyse speech and text inputs.
- Implements reinforcement learning to adjust difficulty levels based on user progress.

Potential Impact:
- Enhances verbal communication and comprehension for individuals with autism.
- Provides a low-cost alternative to traditional therapy, improving access to special education resources.
- Supports educators, therapists, and parents by offering real-time insights into a child’s progress.

The inspiration for this project stems from the growing need for effective, personalized learning solutions for individuals with Autism Spectrum Disorder (ASD). Language acquisition and cognitive processing challenges are among the most significant barriers autistic learners face, often leading to communication difficulties, social isolation, and limited educational progress. Despite the availability of traditional therapy and assistive technologies, these solutions are often costly, time-intensive, and lack adaptability to individual learning needs.

With recent advancements in Artificial Intelligence (AI), particularly in Natural Language Processing (NLP) and Adaptive Learning Models, there is an opportunity to bridge this gap. AI-driven tools can analyze speech patterns, cognitive responses, and learning progress to deliver personalized interventions that cater to the unique linguistic and cognitive abilities of autistic individuals. This project proposes the development of an AI-powered educational application that uses deep learning models, real-time speech analysis, and contextual language training to enhance language development and cognitive skills.

The underlying hypothesis is that an AI-driven, adaptive language processing system can significantly improve verbal communication, comprehension, and cognitive processing for individuals with ASD. The system will succeed because:
1. Personalised AI Models – It adapts to the user’s learning pace and cognitive abilities.
2. Real-time Feedback Mechanism – AI-driven assessment and feedback enhance engagement.
3. Accessibility and Scalability – A cost-effective alternative to traditional therapy, making specialised learning accessible to a broader audience (e.g., language, age group, disabilities,… etc.).
4. Integration of Speech and Text-Based AI Models – Provides a multimodal learning experience that improves comprehension.

Not applicable

The proposed AI-powered application is built on a modular architecture that integrates speech processing, natural language understanding (NLU), and adaptive learning algorithms. The system comprises:
1. Speech and Text Processing Module – Utilizes automatic speech recognition (ASR) and natural language processing (NLP) to analyze spoken and written inputs.
2. AI-Driven Language Model – Employs deep learning (transformer-based models like BERT/GPT) to evaluate linguistic patterns, identify strengths and weaknesses, and generate personalized learning exercises.
3. Adaptive Learning Engine – Uses reinforcement learning to adjust content based on the user’s progress, cognitive abilities, and emotional cues. And
4. Real-time Feedback System – Provides instant assessments and recommendations using AI-powered speech synthesis and visual aids.

Implementation Process & Innovative Features
- Phase 1 (0-3 months): Data collection & preprocessing – Curating speech datasets from autistic individuals, anonymizing data, and training initial NLP models.
- Phase 2 (4-6 months): Prototype development – Implementing speech-to-text models, AI-driven language exercises, and gamification for engagement.
- Phase 3 (7-9 months): Testing & refinement – Conducting user trials with therapists, parents, and educators, refining the AI feedback system.
- Phase 4 (10-12 months): Deployment & optimization – Scaling the model, multi-language support, and enhancing real-time adaptation.

Performance Metrics & Evaluation
- Accuracy of Speech Recognition & Language Comprehension (Target: ≥ 85%)
- User Engagement & Retention Rates (>75% completion of assigned exercises)
- Language Improvement Scores (Measured via standardized tests) e.g., Stanford be een test + parent feedback pre-test
- Reduction in Therapy Costs (Comparative analysis with traditional methods)

This project introduces a highly innovative and AI-driven approach to tackling the linguistic and cognitive development challenges faced by individuals with Autism Spectrum Disorder (ASD). Unlike traditional language therapy methods that rely on structured, one-size-fits-all interventions, our solution leverages adaptive artificial intelligence to provide a personalized, dynamic, and scalable learning experience.

Key Innovations:
1. AI-Driven Personalization:
- Unlike static language training programs, our system adapts in real time to the user’s progress, speech patterns, and cognitive abilities.
- Machine learning models analyze speech fluency, vocabulary usage, and comprehension levels to tailor interventions for each individual.
2.Multimodal Learning Approach:
- Integrates speech processing, text-based NLP, and visual reinforcement techniques, offering a multi-sensory experience that enhances learning for autistic users.
- Uses emotion recognition algorithms to adjust interactions based on user engagement and stress levels.
3. Gamification & Real-Time Feedback:
- The system incorporates game-based learning elements, making the language acquisition process engaging and reducing learning fatigue.
- Provides instant feedback via AI-generated speech synthesis, interactive dialogues, and visual cues, improving retention and motivation.
4. Cost-Effective and Scalable Solution:
- Unlike traditional therapies that require frequent specialist intervention, this AI-powered application can be used independently at home or in classrooms, reducing reliance on human therapists and lowering costs.
- Supports multiple languages and can be adapted to various cultural contexts, increasing accessibility worldwide.

To ensure scalability, sustainability, and long-term impact, this project employs a modular AI architecture and adaptive learning algorithms. These strategies enable the system to meet growing user demand, maintain efficiency, and continuously evolve to address users’ changing needs.

Scalability Strategies:
1. Efficient Machine Learning Models: AI models will be optimised for speed and efficiency, using neural networks and quantisation techniques to reduce computational costs.
2. Multi-Platform Deployment: Designed for mobile, web, and tablet compatibility, ensuring wider accessibility for diverse user demographics. It could also supports multiple languages, enabling expansion into different regions and educational systems.

Sustainability and Long-Term Engagement:
1. Adaptive Learning & Continuous Improvement:
- AI-driven analytics will track user progress and dynamically adjust learning content to keep users engaged.
- Regular model updates based on new research and user feedback ensure continuous improvement.
2. User Retention Strategies:
- Gamification, progress tracking, and community features will encourage long-term usage.
- Integration with schools and therapy centres will ensure sustained adoption in structured educational settings.

This project directly addresses critical social and educational challenges faced by individuals with Autism Spectrum Disorder (ASD) by adapting AI to improve language development, cognitive processing, and social communication. Many autistic individuals struggle with verbal expression and comprehension, limiting their ability to engage in education, employment, and social interactions. By providing a personalized, AI-driven learning tool, this solution enhances independence, confidence, and communication skills, fostering greater inclusion and equity in education and society.

Addressing Social Issues and Promoting Equity
1. Accessibility & Inclusion: The tool is designed to be affordable and accessible, reducing reliance on expensive therapy and making quality language training available to underserved communities. Additionally, multilingual support ensures that individuals from diverse linguistic backgrounds can benefit. Finally, the availability across multiple platforms (mobile, web, tablets), allows users in low-resource settings to access the tool.
2. Bridging the Educational Gap: It enhances special education programs by providing an adaptive, AI-driven supplement to traditional teaching methods. It also empowers teachers, therapists, and parents with real-time insights into a child’s progress, enabling more effective interventions.

Measuring Social Impact

To ensure effectiveness and responsiveness to the community’s needs, the following impact metrics will be tracked:
- Improvement in Communication Skills (measured via standardized language assessment tools).
- User Engagement and Retention Rates (percentage of users actively completing exercises).
- Reduction in Therapy Costs (comparison with traditional therapy expenses).
- Feedback from Educators & Parents (qualitative assessments of real-world usability).

Personal Information Collection Statement (PICS):
1. The personal data collected in this form will be used for activity-organizing, record keeping and reporting only. The collected personal data will be purged within 6 years after the event.
2. Please note that it is obligatory to provide the personal data required.
3. Your personal data collected will be kept by the LTTC and will not be transferred to outside parties.
4. You have the right to request access to and correction of information held by us about you. If you wish to access or correct your personal data, please contact our staff at lttc@eduhk.hk.
5. The University’s Privacy Policy Statement can be access at https://www.eduhk.hk/en/privacy-policy.

• I have read and agree to the competition rules and privacy policy.