Pioneering Core & Applied AI Paradigms
Explore the specialized computational domains supported by AICDS research grants, special interest groups, and symposium tracks.
Generative AI & LLM Orchestration
Investigation into parameter-efficient fine-tuning (PEFT/LoRA), retrieval-augmented generation (RAG) with hybrid lexical-dense vector indexing, multi-agent goal decomposition graphs, and verifiable reasoning pipelines.
- Domain-specific small language model optimization
- Self-reflecting multi-agent tool-use loops
- Hallucination mitigation and attribution verification
High-Velocity Data Engineering
Modern enterprise architectures handling exabyte-scale analytics, distributed vector search mechanisms (HNSW, IVF-PQ), lakehouse query optimization, and real-time streaming feature stores.
- Sub-50ms Approximate Nearest Neighbor indexing
- Distributed graph databases and knowledge reasoning
- Continuous streaming data pipelines with Kafka & Flink
Edge AI & Embedded TinyML
Deploying neural models to low-power edge compute units under 50mW for smart agriculture, remote health sensors, and autonomous drone navigation without continuous internet access.
- Post-training integer quantization (INT8/INT4)
- Hardware-accelerated DSP and NPU runtime engines
- Federated on-device model learning
Quantum Computing & QML
Parameterized quantum circuits, quantum approximate optimization (QAOA), variational quantum eigensolvers (VQE), and hybrid quantum-classical neural network architectures.
- Variational quantum algorithms using Qiskit & PennyLane
- Quantum kernel methods for high-dimensional feature spaces
- Noise mitigation in Noisy Intermediate-Scale Quantum (NISQ) systems