Enterprise AI

Chairpersons welcome and opening remarks. 

As investment in AI continues to grow rapidly, many businesses continue to face the challenge of effectively leveraging AI and maximizing business impact. Although AI has shown incredible power when solving certain problems and driving tremendous value, building enterprise AI excellence is a journey that requires vision, strategy plan, and execution. With the right framework and clear focus on key aspects to drive success including organization, data & analytics capability, technology, and value delivery, the power of AI can be fully unleashed and drive exponential impact across enterprises.

Labeling training data is exhausting—the de facto bottleneck most AI teams face today. Eager to alleviate this pain point of AI development, practitioners have long sought ways to automate this labor-intensive labeling process. Automate too little (e.g., with manual labeling optimizations such as active learning or model-assisted labeling) and the gains are marginal. Automate too much and your model becomes disconnected from the essential human-provided domain knowledge it needs to solve relevant problems. The key to truly transformative (e.g., 10x to 100x) efficiency improvements is to change the interface to labeling altogether, moving from manual labeling collecting individual labels one-by-one to programmatic labeling with labeling functions that capture labeling rationales. The result is a labeling process that is significantly more scalable, adaptable, and governable. In this talk, we review these techniques for automating parts of the labeling process, show how the Snorkel Flow platform integrates them in a unified framework, and share real-world experiences from Fortune 500 companies that have made the transition from manual to programmatic labeling.

• Human-in-the-loop review processes – how can we decide where the buck stops in decision making?  
• Reconfiguring data input to tackle non-ethical AI bias. 
• The potential of labour replacement and corporate practice that can be done to mitigate the risk of job-loss. 

• 5% model, 95% everything else: Three truly comprehensive pillars to evaluate AI ideas for enterprise “viability”
• Tips on building AI teams that get things done
• Agile AI? Waterfall agile AI? Examples of bringing agile principles to AI and machine learning

AI is becoming increasingly important but is only effective with good quality data. Companies often dismiss the importance of this, but what does good data for AI look like? In this session we will discuss how AI powered by the right data is highly effective, and what tools you can use to confidently collect the best quality data from your platforms.

Session focusing on utilizing Cloud and AI in the energy sector. 

To follow soon …

Internal intelligence is one of your companies’ biggest competitive assets. A tremendous amount of information lives in PowerPoints, PDFs, videos, Word Documents and more. This data is the most difficult to extract from, yet the most valuable, as time, money and insight was spent converting data into knowledge. It’s a constant challenge to find the information when needed, especially when information is spread across repositories and geographies. Recent advances in AI-knowledge management platforms are now providing teams the capability to instantly access answers across all sources of internal knowledge and delivering businesses huge advantages from freeing up thousands of hours per year spent on searching for existing information to saving millions of dollars in redundant work.

The global pandemic has changed life – the reduction in face-to-face contact and the pivot to audio has been great for many, but also raises questions around inclusion and diversity. Does today’s technology really understand everyone regardless of age, gender, accent, or location? This talk explores:

• Self-supervised learning with Autonomous Speech Recognition.
• How we can use Autonomous Speech Recognition to tackle inequality and bias in AI.
• The move beyond language accuracy towards voice cohort accuracy will deliver inclusive and diverse speech recognition that makes AI work better anytime, anyplace, anywhere.

To follow soon …

To follow soon …

• NLP (Natural Language Processing) progressions like OpenAI for enhanced user experiences and automated responses.  
• How these advancements are enhancing workforce and service capabilities, as well as increased data literacy.  
• What are the challenges that need to be considered here?  
• Accelerating processes with deep learning and neural networks – looking at the expansion of deep learning at scale across complex tasks.

• End-to-end MLOps platform for 6 different ML frameworks and custom accelerators to transform your ML use cases from concept to production.
• The platform offers 3 steps of MLOps starting with the “Dev” stage where we unit test the ML modules required in your use case.
• This is followed by the “Test” stage where we build ML pipelines with Kubeflow and Airflow.
• Finally the “Prod” stage deploys the ML components into production with model monitoring and CI/CD.
• The platform is supported by code in GitHub with examples.

View day 2 content here:

Applied Data & Analytics

Data Optimisation | Big Data | Intelligent Decision Making | Data Storage | Industry Use-cases 

Transformational AI

Personalisation | Edge AI | Digital Transformation | Future Progressions | Industry Use-case