Scientific Foundation
Built on Temporal Motivation Theory (TMT) and Just-in-Time Adaptive Intervention (JITAI) frameworks.
- Temporal Motivation Theory
- Just-in-Time Interventions
- LinUCB Contextual Bandit
- Zero-shot NLI Classification
- SBERT Semantic Embeddings
Research Gap
Addressing the limitations of static productivity tools that lack behavioral insights and adaptive learning.
- Static vs. Adaptive Learning
- Notification Fatigue
- Multi-layer Monitoring Gap
- Absence of Closed-loop Feedback
- Manual Productivity Tagging
Research Problem
Academic procrastination in digital environments remains inadequately addressed by reactive, static tools.
- Digital Distraction Modeling
- Self-Regulation Failure
- Impulsivity & Delay Discounting
- Task Aversiveness Impact
- Static Intervention Limitations
Research Objectives
Engineering a closed-loop system to monitor, recognize, and mitigate procrastination in real-time.
- Dual-layer Activity Monitoring
- Hybrid AI Pattern Recognition
- Adaptive Task Scheduling
- Smart Intervention Selection
- Behavioral Feedback Integration
Methodology
A sequential data pipeline integrating machine learning models with behavioral theory constructs.
- Progressive 3-Layer Classification
- Hybrid AI (XGBoost, HMM, LSTM)
- SBERT & k-NN Time Prediction
- TMT Context Vector Mapping
- Discounted LinUCB Algorithm
Key Results
Validated through a 6-week deployment with 50 university students in naturalistic settings.
- 78.4% Classification Accuracy
- 67.9% Scheduling MAE Reduction
- 39.4% Intervention Acceptance
- 76.9% Learned System Silence
- Significant Adaptive Learning
Project Proposal
Progress Presentation I
Progress Presentation II
Development Phase
Project Velocity
Technical Documents
Topic Assessment
Project Charter
Project Proposal
Checklist I
Checklist II
Checklist III
Research Paper
Final Report
Poster
Feedback
What Our Users Say
"The adaptive notifications are a game changer. Unlike other apps, it actually learns when I'm most likely to procrastinate."
Engineering Student
SLIIT Participant
"Seeing my TMT scores in real-time helped me understand my own patterns. The 'Reframe' interventions were my favorite."
Computing Undergraduate
6-Week Study User
"The task decomposition actually makes big projects feel manageable. It's like having a productivity coach built-in."
Researcher
Alpha Tester
Amaratunge A.
Behavioral Monitoring
Dual-source tracking & three-layer activity classification pipeline.
it22351586@my.sliit.lk
Vilochana A.G.B
Pattern Recognition
Hybrid AI models for behavioral pattern detection & risk scoring.
it22114808@my.sliit.lk
Jayasundara S.M.A.V
Task Prioritization
LLM-based task decomposition & adaptive duration prediction.
it22352576@my.sliit.lk
Jayasinghe N.P.
Smart Interventions
Contextual bandit system for personalized interventions.
it22202468@my.sliit.lk