Big data & AI

We provide big data analytics and AI solutions to meet the intelligence requirements of clients' biomedical and healthcare projects. We have solid skills and deep hands-on experience in implementing libSVM, Apache Spark, and TensorFlow. Moreover, we are experts in tackling full stages of machine learning projects including data collection, extraction and normalization, training and testing machine learning models, cross-validation and fine-tuning, rule generation, and deploying real-time detection/prediction applications.

Key features and benefits:

• We carefully examine each variable/feature from raw data to ensure that each variable is correctly transformed to a predictor.
• Our solutions can be deployed on SGE/SGI clusters or cloud systems as chosen by clients.
• We can implement Hadoop HDFS or in-memory frameworks like Spark, depending on computing resources and project goals.
• We deliver both machine learning models with overfitting fully excluded, and regression rules between predictors and outcomes.

As an example, we successfully constructed a genetic variant database which hosted 7000 whole-genome sequencing samples and enabled genotype querying and allele frequency computing within minutes, based on a generic MapReduce framework within-memory and indexing technologies deployed on SGI UV 2000. Learn how