Deep Learning uses artificial neural networks, it is inspired from human brain. Deep learning, a term that alludes to the many (deep) layers within neural networks. Deep learning has powered many of the recent advances in AI.
What is Gen AI?
Generative AI is a subset of Deep learning, it uses artificial neural network, can process both labelled and unlabelled data using supervised, unsupervised and semi supervised methods. Generative Deep learning model, learn patterns in unstructured content, generate new data that is similar to data it was trained on.
Generative AI, foundation models and Transformers - a love story
The underlying model that enables generative AI to work is called a foundation model. Transformers are key components of foundation models. A transformer is a artificial neural network that is trained using deep learning. However, some characteristics set foundation models apart from previous generations of deep learning models. To start, they can be trained on extremely large and varied sets of unstructured data.
What CEO wants from their teams
The organisational requirements can be classified into a few broader categories depending on the business imperatives and the use cases.
Consuming GenAI outcomes as they are offered from any 3rd party hosted GenAI SaaS solution. This requires the least change in the tech culture apart from just choosing the right solution and integrating it with existing tech stack.
Accessing foundational model through an API, building post processing layers, followed with providing user interface for exposing the customised outcomes will need a small team and bit of integration. This work still uses existing foundational models as they are.
Enriching foundational model with proprietary, labeled data set requires to fine tune the model to provide enhanced, niche outcomes to customers. This requires setting up datastore, infrastructure cost, expertise human capital and model maintenance cost to be considered.
Accelerating the pace of research may need to train a foundation model from scratch. Training such a model will require large data. To create a proprietary data, we need dedicated human capital for data engineering, data science and tech infrastructure.
Briefing the considerations of a CEO
Generative AI requires a more deliberate approach given its unique risk considerations and the ability of foundation models to underpin multiple use cases across an organisation.
- Eliminate risks.
- Balancing risk and value creation.
- Cost of pursuing GenAI may vary depending on the category of business imperatives.