Our Core Capabilities
ML, AI and Predictive Analytics
People often use the terms AI and ML interchangeably. The two are very closely linked as ML provides the backbone for the modern methods that drive AIs. ML algorithms learn extremely complex patterns from data, sometimes giving the impression of human cognitive abilities or ‘intelligence’, however fundamentally serve the function of making predictions from data.
At Gwendra AI we are concerned with the practical application of these technologies. In this context, the models make predictions. Predictions then inform judgement and decisions leading to actions. We consider the full Predictive Analytics process, looking at data, predictive models and the subsequent response and actions.
ML techniques require large volumes of well labelled data to learn from. Many recent advances have capitalised on the huge amounts of data generated by billions of people accessible on the Internet: so called Internet AI.
Business information typically comprises smaller, information rich data. Often this data is unstructured and weakly labelled which can be a problem when applying current ML capabilities. At Gwendra AI we specialise in Business AI; that is helping organisations gain competitive advantage from current ML based models with their their unique, but ‘difficult’ proprietary data sets.
Natural Language Processing
NLP forms a core part of many of our models and applications. Tools to process textual information, drawing out facts, meaning and sentiment give us the ability to access the vast wealth of data outside of traditional databases of structured data.
We use state of the art neural network based NLP tools, coupled with proprietary deep learning capabilities on a high performance computing platform to create bespoke NLP processing pipelines tailored and tuned to the client’s individual data, use cases and requirements.
Deep Learning improves the capabilities of Artificial Neural Network (ANN) technologies and is at the heart of many of our models. ANNs are inspired by our understanding of how the neurological structures in our brains and visual cortexes work. Recent developments have opened up the capabilities of these models, comprising the biggest breakthrough in ML for years.
Statistics, Time Series Analysis and Statistical Methods
While neural network based models and NLP feature strongly in our solutions, descriptive statistics form an essential part of our upfront data analysis and statistical analysis of model performance.
We also use time series analysis and modelling to identify patterns in seasonal data and sometimes use this to impute data, increasing datasets for ML applications. Bayes and other techniques, including XG Boost are used for early model prototypes and when datasets are simply too challenging for other methods.
Gwendra AI Lab
Model development and production applications have very different characteristics. Production run-time environments must be stable, reliable, have predictable performance and high availability.
Data analysis, model experimentation and development require flexibility to give the analyst the opportunity to quickly turn around prototypes. This is not conducive to a controlled production environment. Also, model training and evaluation, especially with deep learning models is computationally intensive. These intensive operations require GPU and other high-performance computing tools and can accumulate significant cost in cloud-based platforms.
We overcome these competing requirements by implementing two environments: a cloud based production environment and a separate open, flexible and high-performance on-premise computing environment for model development, training and evaluation.