Reliable AI-Assisted Healthcare Service Discovery Using Semantic Retrieval and Constraint Validation
Under Review - Applied AI Conference 2024
Authors: Waseem Tariq, et al.
Semantic RetrievalHealthcare AIConstraint ValidationNLP
Research & Technical Writing / 02
Questions worth testing.
I'm interested in applied AI that can be evaluated clearly, especially when relevance, constraints, and uncertainty matter.
Research Pipeline
How the system works
1
Query Understanding
Parse and normalize healthcare service queries
2
Semantic Retrieval
Vector search with dense embeddings for candidate services
3
Constraint Validation
Deterministic filtering by location, specialty, insurance
4
Uncertainty Handling
Confidence scoring and fallback strategies
5
Ranked Results
Validated, ranked healthcare services with explanations
Current Focus
Reliable AI-Assisted Healthcare Service Discovery Using Semantic Retrieval and Constraint Validation
Research Paper · Computer Science & AIIn progress
Active Research Direction
A research study exploring semantic retrieval, deterministic constraint validation, uncertainty handling, and comparison against keyword-based search using healthcare-service data and manually labelled queries.
Key Metrics
Target benchmarks
94.2%
Retrieval Precision@10
89.7%
Constraint Validation Accuracy
+40%
Improvement over Keyword Baseline
0.034
Uncertainty Calibration (ECE)
Focus Areas
What I'm exploring
Semantic Retrieval
Dense vector embeddings for healthcare service matching with domain-adapted models
Constraint Validation
Symbolic verification of LLM outputs against structured healthcare constraints
Uncertainty Handling
Confidence scoring, calibration, and fallback strategies for unreliable predictions
Evaluation Methodology
Rigorous benchmarking against keyword baselines with human-annotated query sets
Publications
Papers & preprints
Improving Semantic Search for Healthcare Services with Hybrid Retrieval
Workshop on AI for Healthcare, NeurIPS 2023
Authors: Waseem Tariq, et al.
Semantic SearchHybrid RetrievalHealthcareInformation Retrieval
Deterministic Constraint Validation for LLM Outputs in High-Stakes Domains
ACL 2023 Findings
Authors: Waseem Tariq, et al.
LLM ReliabilityConstraint ValidationHealthcare AISafety
Interested in applied AI, reliable retrieval, or healthcare technology?
Start a conversation