AMDM logo 2AMDM Brain

The Accelerated Materials Development for Manufacturing (AMDM) programme in A*STAR aims to speed up the pace of materials innovation by utilizing Machine Learning (ML) and Artificial Intelligence (AI) with High Performance Computing (HPC) and Robotics/ Automation.  We hope to achieve this by closing the materials innovation loop, through linking the three pillars of discovery: 1) Synthesis 2) Process optimization and 3) Rapid characterization and diagnosis. Our programme is uniquely positioned to leverage upon Singapore and A*STAR’s infrastructure and expertise in High Performance Computing, Autonomous Research and Machine Learning. 

The AMDM programme is currently focusing on 5 different fields of materials science:   

  1. Flow synthesis 
  2. Formulation and wet chemistry
  3. Thermal and thermoelectrics
  4. IC defect detection
  5. Soft magnetic materials and high entropy alloys


Capabilities & Collaboration Opportunities

We have a group of dedicated scientists, supported by deep materials knowledge and state of the art equipment to excel at this forefront platform methodology. The AMDM methodology can potentially permeate through a wide variety of different industries spanning from materials synthesis to functional properties and system-level design. We have the in-house expertise and specifically interested in the following speciality chemicals and materials sectors:


Accelerated Materials Development for Manufacturing Programme

We seek and welcome various industrial or academic collaboration. In general, we suggest 3 possible ways where external parties can leverage on this platform methodology. 

  1. Data 

    Share your data with us and we can help you to optimize your process, design better materials or identify system bottlenecks

    Our value add – Machine Learning algorithms catered to materials datasets, materials science descriptors

  2. Material Design and testing in A*STAR

    (e.g.) Inverse design of small molecules, polymers, hybrid materials

    Our value add – 20+ years of materials science experience, synthesis and formulation screening with experiments and supported by theoretical modelling

  3. High Throughput Experiments

    >10x increase in the speed of synthesis and characterization

    Our value add – Completely automated experiments enabled by machine learning algorithms

Highlights & Achievements

  • Inverse Design of polymers 

AMDM Polymer Architecture

The team has made 20 new polymers in the lab that was completely designed by a computer. The machine was able to guide the synthesis of the polymer for a specific desired property 

npj Computational Materials, volume 5, Article number: 73 (2019) 


  • Accelerated testing of functional materials (thermoelectrics and photovoltaics)

    AMDM Bayesian Inference

    The team has developed a Method to Infer Properties of Thermoelectric (TE) Materials from Simple Power Versus Load Data Using Bayesian Machine Learning.

  • Accelerated characterization of Materials Structure

AMDM XRD Pattern

The team has developed a fast and interpretable classification of small X-ray diffraction datasets using data augmentation and deep neural networks.

Oviedo et. al. npj Computational Materials, volume 5, Article number: 60 (2019) 



J.P. Correa Baena et al., Joule 2, 1410–1420 (2018)


Dr. Kedar Hippalgaonkar,

Tonio Buonassisi –

Jatin Kumar,

Dr. Lim Yee Fun,


We welcome queries and collaboration partners to work together on leveraging our automation and machine learning platform for process optimization and new materials discovery