Quantum Computing for Quantum Chemistry and Material Discovery

Fri, 29 May 2026 00:00:00 +0000

Project Mission

This project develops hybrid quantum-AI methods for quantum chemistry, drug discovery, and semiconductor materials design.
We focus on quantum neural networks and quantum-compatible learning architectures that can represent molecular interactions, electronic structures, and materials properties in ways that are
difficult to capture with conventional AI models.

The long-term goal is to connect AI-driven screening with quantum computing resources to enable more accurate and scalable discovery
of drug-like molecules and next-generation semiconductor materials, especially perovskites and optoelectronic materials.

Scientific Motivation

Drug and materials discovery often requires modeling quantum-level phenomena such as electron correlation, molecular binding, charge transport, band gaps, excited states, and defect behavior. These properties are central to molecular design and semiconductor discovery, but they are expensive to predict across large chemical spaces.

Quantum-AI hybrid methods provide a practical pathway by combining the scalability of machine learning with quantum-inspired or quantum-compatible representations. Rather than replacing classical simulation, this project uses quantum neural networks as a bridge between data-driven prediction and future quantum-enhanced computation.

Research Approach

Quantum Neural Networks

Parameterized Quantum Circuits: Designing trainable quantum circuit models for molecular and materials property prediction.
Quantum Feature Encoding: Representing molecular descriptors, features, and structure information in quantum-compatible forms.
Hybrid Quantum-Classical Learning: Combining quantum neural network layers with classical ML/DL models.
Hardware-Compatible Architectures: Developing noise-aware
QNN models that can be deployed on real quantum processors.

Quantum Chemistry for Drug Discovery

Molecular Property Prediction: Predicting stability, reactivity, toxicity features, and electronic properties of candidate molecules.
Drug-Target Interaction Modeling: Learning representations relevant to binding affinity, selectivity, and interaction patterns.
Quantum Chemistry-Informed Screening: Prioritizing drug-like compounds using electronic-structure-aware descriptors.
Reaction and Conformation Modeling: Studying molecular configurations and reaction-relevant quantum effects.

Perovskite Semiconductor Material Discovery

Band Gap Prediction: Modeling electronic properties relevant to photovoltaic and optoelectronic performance.
Perovskite Materials Design: Exploring halide, hybrid, and lead-free perovskites for stable and efficient semiconductor applications.
Defect and Interface Modeling: Understanding how atomic-scale defects influence charge transport, stability, and device behavior.
Qunatum AI-Guided Candidate Selection: Screening materials for solar cells, LEDs, photodetectors, and next-generation devices.

Hybrid System Design

The current framework is designed as a hybrid discovery pipeline. Classical AI models generate, encode, and prioritize candidate molecules, while quantum neural network components learn quantum-compatible representations for property prediction. In future stages, quantum computing modules will be introduced to support electronic-structure estimation, molecular simulation, and quantum-enhanced learning.

This structure allows the project to begin with simulator-based QNN development while keeping the model architecture compatible with future quantum hardware processors.

Current Implementation

At the current stage, this project focuses on quantum neural network models without direct use of quantum computing hardware. The initial implementation uses quantum circuit simulators and hybrid neural architectures to evaluate how QNN layers can encode molecular and semiconductor materials information.

This stage is designed to establish model architecture, data representation, training stability, and benchmark performance before moving toward hardware-based quantum computation.

Future Research Directions

Future work will integrate actual quantum computing resources into the discovery pipeline. The focus will be on making quantum neural networks more compatible with real quantum hardware by improving circuit depth, noise robustness, qubit encoding, and hybrid optimization.

The long-term direction is to build a quantum-AI discovery workflow in which classical AI models identify promising molecular and materials candidates, while quantum computing resources provide more direct support for electronic-structure modeling, molecular simulation, and quantum-enhanced property prediction.

Keynote: AI for Scientific Discovery

Sun, 15 Dec 2024 14:00:00 +0000

About the Talk

Prof. Jane Smith will deliver the opening keynote at the prestigious International AI in Science Symposium, highlighting our lab’s pioneering work in computational biology and drug discovery.

Key Topics

AI-Powered Drug Discovery

Machine learning platforms for molecular design
Accelerating clinical translation from years to months
Real-world case studies and clinical outcomes

Protein Structure Prediction

DeepFold architecture and breakthrough results
Comparison with AlphaFold and novel improvements
Applications in understanding disease mechanisms

Future of AI in Science

Emerging opportunities in materials science and climate research
Challenges in model interpretability and validation
Building interdisciplinary research partnerships

Expected Impact

This keynote will position our lab as a thought leader in AI-driven scientific research and attract potential collaborators from leading institutions worldwide.

Conference Details

Venue: Stanford University Memorial Auditorium
Date: December 15, 2024
Time: 2:00-3:00 PM PST
Audience: 500+ researchers, industry leaders, and policymakers
Format: 45-minute talk + 15-minute Q&A

Media Coverage

The talk will be livestreamed and recorded, with proceedings published in the Journal of AI in Science special issue.

Drug Discovery | Research Lab