Advanced AI: ML Case Study by Ardigen

Ardigen provides advanced AI, bioinformatics and research expertise to help get the most value out of your data. By applying Ardigen’s tailored AI methods, you can uncover and validate novel hypotheses, automate manual workflows and take advantage of the power of machine learning at scale. With years of experience and synergy of expertise, Ardigen drives digital transformation for biotech and pharma companies through advanced quantitative and AI-based approaches. Learn more about their general services here.

Case Study: End-to-end parter for ML project

Overview:

Working with a big pharma client, Ardigen enhanced a large knowledge graph model integrating over 20 data sources, streamlined the evaluation process for novel scientific hypotheses, and developed a user-friendly interface for the efficient management of machine learning experiments.

Collaboration scheme:

Results:

Ardigen achived an ML framework for dataset handling and validation, improved compuational efficiency and model accuracy, a user-friendly interface for interaction with ML engine i.e.: uploading data, traning models, and deployment, as well as a full workflow automation leading to significantly increased utilization by the client’s scientists.

10 times faster: Ardigen accelerated their client’s model more than 10 times and automated model deployment with 10+M parameters & 10+ Gb data input.

10-15% increase in accuracy: Ardigen improved classification accuracy by 10-15% points against the initial DL-based method by carrying out a full benchmark and validation with state-of-the-art models.

Method:

Aridgen combined a rigorous ML methodology benchmark, graph-based ML algorithms, end-to-end ML-Ops framework based on AWS Sagemaker and AWS Cloud and a combination of efficient and robust front-end and back-end technologies.

Aridgen’s AI-Enabled Services

• Design and deployment of custom machine learning (ML) or deep learning (DL) algorithms and models. Learn more here.
• End-2-end analysis of phenotypic High Content Imaging (e.g. Cell Painting) for drug discovery research. Learn more here.
• Design and optimization of new drug candidates using Generative AI or Reinforcement Learning algorithms.Learn more here.
• Multi-omics and multi-modal approaches for early drug discovery and clinical-stage biomarkers identification. Learn more here.
• Engineering services to deploy, support and maintain (large) AI models on client’s infrastructure or cloud environments. Learn more here.
• Computer vision algorithms for biomedical imaging, diagnosis or molecular biology. Learn more here.
• Refinement and customization of large sequence- or structure-based models for protein design and optimization of biological therapeutics. Learn more here.
• Deployment of large language models (LLM) for scientific purposes or optimisation organizational processes
• Automated data processing from various sources using ML-, DL- or LLM-based algorithms
• Implementation of AI-Lab-Loop workflows based on coupling AI models with experimental feedback and active learning approaches

Aridgen’s expertise is supported by a significant publication record across drug discovery areas: PubMed.