Quantum AI

 

What do you get when you combine the deep learning AI field with quantum computing?  You get into the new realm of increasing computational power exponentially that might one day solve the current unsolvable problems.

As of today, cell phones, computers, and supercomputers operate in a binary system where the classical bit stores information as zero or one.  In quantum computing, a quantum bit (or qubit) can store both zero and one at the same time, a property called superposition.  This special property can process many calculations simultaneously, which allows quantum computing to solve complex problems more efficiently.

qAIntum.ai is a company at the forefront of building large language models (LLMs) with quantum computing. The company’s cutting-edge innovation is the Quantum Transformer Architecture where quantum neural networks (QNNs) are integrated with traditional transformer models.  The company’s objective is to build Quantum Large Language Models (QLLMs) with a potential creation of more sophisticated and accurate language models.

                                                                                                       Image:  qAIntum.ai

Then, there’s the Quantum AI group at Google that has been working on quantum computing for almost two decades.  This group has achieved quantum supremacy in 2019 where computation took just 200 seconds that would have taken the most powerful supercomputers thousands of years to calculate.  In 2023, the group achieved another milestone quantum error correction by showing that it’s possible to reduce errors by increasing the number of qubits.

Whether quantum computing AI is hype or reality, I have to give respect to the brilliant scientists and researchers who dedicate their professional lives to work on these high risk theoretical concepts.  These pioneering researches might not have clear real-world business applications initially, but if succeeded could advance the technological progress for humanity.

AI Planet

 

As AI continues to advance and integrate into the many facets of personal and professional life, it’s important to have some knowledge of this deep learning field in order to stay current.  However, for busy professionals who juggle between life and work, the dilemma is how to carve out the free time for learning.

Introducing the 5-week Data Science bootcamp from AI Planet, a compact bootcamp perfect for busy professionals.  This course is designed to give beginners a good foundation of data science with practical exercises.

AI Planet’s benevolent intention is to democratize data science learning so it provides this bootcamp and other data science courses for free. Taking this bootcamp might not turn you into an AI/ML engineer overnight, but it gives you an excellent foundation from which you can build upon your data science knowledge further.  Good luck!

LQM

Since the initial release of ChatGPT-3.5 in November 2022, Large Language Models (LLMs) have received huge attention as the generative AI model.  Subsequent LLMs such as OpenAI GPT-4o, Meta Llama 3.1, or Anthropic Claude 3 Sonnet continue to generate high interest.

There’s another trend that the next AI wave will include Large Quantitative Models (LQMs), which are designed to handle complex data sets and capture the intricacy of quantitative relationship that LLMs can’t.  You would use LLMs which focus on language tasks to summarize documents, but to discover the next cancer drug you would need LQMs.  LQMs utilize advanced computational techniques and data analysis to make predictions, identify trends, and optimize outputs.  The impact of LQMs will be enormous because these AI models will accelerate new discovery in pharmaceutical drugs, chemical compounds, financial models, weather patterns etc.

Advantages of LQMs:

LQMs have many advantages over LLMs and conventional predictive AI models:

1)  Precision:  LQMs excel in tasks that require numerical precision and the ability to analyze vast quantitative data and to model complex mathematical relationships.

2)  Interpretability:  LQMs address the “black box” nature of AI models by offering enhanced interpretability.

3)  Flexibility:  LQMs can be fine-tuned for specific quantitative tasks, providing powerful tools for data-driven decision-making.

4)  Robustness:  LQMs can understand hidden patterns in data and create new synthetic data to add to the limited historical data, making forecasts and predictions more reliable and accurate.

Overall, LQMs promise to usher in the next wave of deep knowledge exploration and discovery.  LQMs will be critical in fields such as finance, healthcare, energy, supply chain management, climate modeling, manufacturing, and retail.  Some examples of LQMs being used in the industry include Flatiron Health, GE Predix, or Paige.  While LQMs also have challenges, their potential benefits in improving decision-making and optimizing outcomes are substantial.

Momentum AI

 

I had the opportunity to attend Reuters’ Momentum AI conference in San Jose this week.  This summit provided a platform for business executives to showcase their AI thought leadership, challenges, and breakthroughs.  

Some highlights from the conference:

Five Stages of AI Progression

1)  Conversational AI:  AI systems that can engage in natural language conversations with humans such as chatbots or virtual assistants.

2)  Reasoning AI:  AI systems capable of logical reasoning and problem-solving such as advanced decision-making algorithms.

3)  Autonomous AI:  AI systems that can operate independently without human intervention such as self-driving cars and autonomous drones.

4)  Adaptive AI:  AI systems that can continuously learn and improve from their experiences based on feedback such as recommendation systems and adaptive learning platforms.

5)  Organizational AI:  AI systems integrated into the organizational structure of businesses, transforming operations, decision-making processes, and strategic planning. These systems can optimize workflows, enhance productivity, and drive innovation across departments. 

 

Many thanks to Reuters Events for organizing this venue for professionals to share AI knowledge and foster business relationships.

Model Evaluations

 

Model evaluations (or ‘evals’) are the steps that use different metrics to evaluate an AI/ML model’s performance.  These evals are important for assessing performance, measuring effectiveness, ensuring safety, and improving overall reliability of the model.  Model evals can be done in two ways:

• Offline:  Evaluated after model training or continuous retraining. 
• Online:  Evaluated in production as part of model monitoring.

Let’s ask GPT-4 to summarize the key aspects of evals.

For supervised learning, some popular metrics for classification model evaluation include:

• Confusion Matrix provides a summary of prediction results on a classification problem.  
• Accuracy measures how often the classifier makes the correct predictions.  It’s the ratio of number of correct predictions over total number of predictions.
• Precision measures the accuracy of the positive predictions made by the model.  It’s the ratio of true positive predictions to the total number of positive predictions (both true positives and false positives).  A high precision indicates that the classifier has a low false positive rate, meaning it rarely classifies negative instances as positive.  Precision is an important metric when you want to be very sure of your prediction or when the cost of false positives is high.  For example, in medical diagnostics for a rare disease, a high precision ensures that healthy individuals are rarely misdiagnosed as having the disease.

• Recall (aka sensitivity or true positive rate) measures the model’s ability to identify all positive instances.  It’s the ratio of true positive predictions to the total number of actual positives (the sum of true positives and false negatives).  A high recall indicates that the classifier successfully identifies most of the positive instances.  Recall is often traded off with precision. While recall focuses on identifying all positive instances, precision focuses on the accuracy of positive predictions.  Improving recall can sometimes lead to a decrease in precision and vice versa.

• F1 Score provides a balanced measure of a model’s performance by combining precision and recall into a single value.  

Let's use a spam email classification as an example:

1)  True Positives (TP): 70 emails correctly identified as spam.

2)  False Positives (FP): 10 emails incorrectly identified as spam (they were actually not spam).

3)  False Negatives (FN): 5 emails incorrectly identified as not spam (they were actually spam).

4)  True Negatives (TN): 100 emails correctly identified as not spam.

Accuracy =  (TP + TN) / (TP + TN + FP + FN)  

=  (70 + 100) / (70 + 100 + 10 + 5) =  91.9%

Precision =  TP / (TP + FP)  =  70 / (70 + 10) =  87.5%

Recall =  TP / (TP + FN)  =  70 / (70 + 5) =  93.3%

F1 Score =  2 * (Precision * Recall) / (Precision + Recall)

=  2 * (0.875 * 0.933) / (0.875 + 0.933)  =  90.2%

The choice between precision, recall, and F1 score depends on the specific context and objectives:

• If false positives are more costly, prioritize precision.
• If false negatives are more costly, prioritize recall.
• If a balance is needed, the F1 score is a good metric to use.

The F1 score provides a single, comprehensive measure of a model's accuracy by balancing precision and recall, making it a valuable tool for evaluating and comparing classification models.

KissanAI

 

Introducing KissanAI, the generative AI agriculture platform that provides farmers with personalised, voice-based assistance for all their agricultural needs.  KissanAI has a simple, user-friendly interface, making it easy for farmers at any levels of education to use.

Although KissanAI was created in India, however, it was not built for Indian farmers only.  KissanAI's mission is to empower farmers worldwide with the latest technology and knowledge to improve their agricultural practices and increase their yields.  

It’s encouraging to see that the people at KissanAI use the latest AI technology to help the often underserved and overlooked farmers in India and around the world.  Even if you're not a farmer, try it out because you might benefit from some useful tips on how to make your garden blooming with flowers!

Kubernetes

 

Kubernetes (or K8s) is an open-source, software platform that manages clusters of containerized applications and services across private, public, and hybrid cloud environments .  This platform acts like an orchestration system that allows users to deploy, scale, schedule, and manage containerized applications seamlessly across clusters of nodes. 

Let’s ask GPT-4 to explain Kubernetes in a simple way.

Developed by Google and now maintained by the Cloud Native Computing Foundation (part of the Linux Foundation), Kubernetes is often described as the "operating system for the cloud."  Here’re the functions of Kubernetes and what they can do:

About 60% of organizations have adopted Kubernetes and many of these companies are in the AI arena.  Here're some reasons why Kubernetes is popular: