May 25, 2021
The Antidote To The Hype, Noise, And Spin Of Artificial Intelligence
By Jason Mars
May 25, 2021
The Antidote To The Hype, Noise, And Spin Of Artificial Intelligence
What went wrong with artificial intelligence? This transformative technology was supposed to change everything. I’ve seen first-hand the incredible potential it has—both as a professor of computer science at the University of Michigan and as the founder of Clinc, ZeroShotBot, Myca.ai, a non-profit called ImpactfulAI, and several other AI-focused companies.
So, why has it devolved into overhyped solutions, marketing noise, and an endless spin of the same, tired ideas? Into poor user experiences, embarrassing bugs, and countless other misfires?
The answer is pretty clear when you consider how every business has been told it needs artificial intelligence to stay competitive. This mad dash is symbolic of the gold rush, as companies push and pull to be early adopters—to scrape every last dollar out of their ROI. Add to that the misconceptions about what it can do, the ebb and flow of innovation vs. standard techniques, the grandiose promises, the marketability of AI, and it becomes clear how we got here.
It makes me sad to see AI reduced to a gimmick. To be clear, I’m not saying AI doesn’t have an important role to play. It will define the future of technology in many ways. The challenge is looking beyond the noise.
What’s the most important rule of AI? Don’t believe it unless you can see and feel it.
Why do I think this is the most important mental model? The magic of AI still exists, there are places where innovation still occurs, and when it does, the results are undeniable. Having said that, you can’t escape the noise, the hype, the big promises.
Simple, purpose-built AI solutions have transformed many industries. AI is being used in healthcare to detect breast cancer, in agriculture for crop yield forecasting, in autonomous driving to improve safety. These solutions use deep learning and reasoning to draw conclusions from billions of analyzed pixels. There’s no denying these use cases. They’re clear as you can actually see it in action and see it working well.
Trusting this type of intuition must be applied in all realms.
Throughout my experience creating novel conversational AI technologies, I know the power of an unforgettable experience. When it’s real, you know it. It only take a few minutes of interaction to tell if another human is intelligent, and similarly, you know right away if a conversational AI is intelligent from actually interacting with it. You have to look past the canned experiences, the lofty promises, and see what AI looks like in practice—within your industry or use case.
And if something sounds fake or unbelievable? It probably is. Trust your senses, they will guide you through the noise.
Maybe you beat the odds and found that perfect AI solution. It can happen, right? Take a step back and think about the bigger picture. How will you apply that solution to your needs?
A promising demo isn’t everything. You still have to adapt that AI for your use case, train it, deploy it, and improve it. The more niche and customized your use case is, the harder it will be to realize the AI quality demo’d into reality in your environment. When the quality of your AI requires specific training to your use case, production-grade AI is extremely complex and often requires a dedicated team of experts in machine learning, computer and data science, and training specialists. Each layer adds more complexity, making your solution more expensive, brittle, and likely to fail.
As chronicled through my journey as CEO of Clinc, I saw countless companies spend millions trying to create, configure, and train virtual assistants, only to fail. The learning curve is steeper than ever, and the stakes are even higher.
So, how can you successfully navigate the world of AI? It starts with asking the right questions, things like:
And even if you know the answers to these questions, that same demo experience you saw may be untenable if you have to train the AI yourself.
You must be reasonable about the logistics of making AI possible. Be ready for these costs: engineers to work it, support to keep it running, and training specialists (data scientists / ML experts) to improve it.
Next, ask yourself how it ties into mission criticality. Can you afford for it to fail? What’s at stake if your AI spectacularly fails? What will happen if you change the model’s task?
AI is some of the most complex technology on the planet. Getting it right means defining your expectations and knowing your limitations.
Let me start by saying we are already in an AI revolution, thanks to advancements in deep learning, which uses data to model the way our brain’s neural network works. The catalysts for this initial success include the availability of data, advancements in deep learning models, and innovations in computing.
Despite this, not all AI problems can be solved by advancements in neural networks. Many companies may claim to use next-generation AI, but more often than not, it’s just noise in the AI hype cycle.
Here’s what I can tell you. The biggest advancements are occurring in areas where they use deep learning techniques and data to train a system, such as in Natural Language Processing (NLP) or computer vision.
Think about it like this. If we see large amounts of data being used to extract patterns, that’s a direct representation of the AI revolution. This type of approach being the basis of new products like Myca.ai is where AI is leveraged in a transformational way.
So, where are things going wrong? Most companies are using old techniques to latch onto the AI hype cycle. Think about early chatbots and the frustrating user experiences they offered. These solutions used the old Stanford NLP library and similar classical computational linguistic approach that leveraged grammar, nouns, synonyms, dictionaries, and other linguistic mechanics to derive patterns.
The problem? This is the wrong approach in modern times. You can’t expect to innovate if you rely on antiquated techniques.
Now for the big question: how can you see through the noise and see if an AI solution is legitimate? I recommend you learn a selection of latest buzzwords to see if they apply to a given technology.
If they use computational linguistics, regression models, or decision trees, it’s antiquated.
If they use neural networks, transfer learning, adversarial networks, or attention models, it’s current.
You don’t need to understand how they work theoretically. Your focus is knowing what buzzwords to spot and inform yourself of trends through projects like ImpactfulAI. Look for things like convolutional neural networks, transformers, attention models, GANs to quickly identify if the underlying technology is part of the AI revolution.
2020 was a milestone year for the scientific community. A little something called Generative Pre-trained Transformer 3 (GPT-3) was developed by the OpenAI project. The language model it represents is based on 175 billion parameters and is more accurate than anything we’ve ever seen. For context, the older GPT-2 used 1.5 billion parameters.
This model was inspired by recent work on transfer learning, which was first popularized by the Bidirectional Encoder from Transformers (BERT) model, and is built on the belief that you can train an AI model really well once with massive amounts of data (say the entire internet) then use significantly less or no training data for a new task.
This transformational work popularized a new philosophy toward deep learning models as “few-shot learners,” “one-shot learning” and “zero-shot learning,” meaning only a few, a single, or no training examples are needed for the model to perform a completely new task.
Let that sink in for a moment. For the first time, we may be able to create a conversational AI for new types of problems without any training. With the AI philosophy introduced by GPT-3, you can ask any question and receive an incredibly accurate answer without the need of training. One of my next major endeavors is to introduce the first commercialization of such an approach through the development of Zero Shot Bot to revolutionize the conversational AI chatbot space, and the performance of this technology is breath taking.
I firmly believe that GPT-3 and now Zero Shot Bot serves as a bell weather of the next big game-changer for the next decade. It’s not a matter of if, but when. This Zero Shot to Few Shot approach is the answer to the user experience problems created by other antiquated AI technologies. In the context of the Internet, it can solve a host of interesting problems and doesn’t require any training.
And Microsoft agrees. They signed a staggering $1 billion licensing agreement with OpenAI largely due to how impressive this model is.
This philosophy is now in the air as of just a year ago. Beyond GPT-3 and Zero Shot Bot, no products exist yet to my knowledge. But mark my words, industry-changing technology and commercializations of this Zero Shot approach is coming.
Zero Shot Bot and other platforms that use zero-shot learning removes the hardest part about deep learning AI—the training.
If you ask me, that’s the magic spark of innovation that commercial AI has been missing for years.
May 18, 2021
What’s In Store For The Next Generation Of AI? The Jaseci Perspective
By Jason Mars
May 18, 2021
What’s In Store For The Next Generation Of AI? The Jaseci Perspective
And the more you work with these popular programming languages, the quicker you discover the true meaning of compromise. That’s not to say these languages aren’t great at what they do, because they are.
The problem is that AI is extremely complex. The commercial prevalence of AI is so new that the existing languages aren’t suited to the unique challenges AI presents.
It’s important to understand the key differences between the academic sphere and the business world. AI exists in both, but the mental framework is vastly different. In academics, the focus is on the science and theory of AI, with some degree of the empirical. In business, it’s all about what AI looks like in practice and the tangible outcomes it provides.
As a professor of computer science at the University of Michigan, I’ve worked on the cutting edge of AI. I’ve seen where current technology falls short when it comes to AI processing, building AI-driven ecosystems, and using cloud technology to run large-scale AI. I’ve seen firsthand the stark differences between the science behind it all to the art of what is possible.
That alone wasn’t enough. I knew I needed to see what commercial AI looked like. When I founded Clinc, that experience quickly became a five-year crash course for commercial AI through its success under my leadership. I saw what real-world challenges existed for large-scale AI systems, the processing requirements for enterprise AI, and the millions of queries being made.
My conclusion? The current computational model and programming paradigm is ill-suited for the emerging set of problems in the AI sphere. A new computational model and programming language are desperately needed. The way we articulate solutions to problems in computer science has not fundamentally changed in decades and we’re hitting a new ceiling with the unique complexities that come with building sophisticated AI solutions.
That’s why I created Jaseci, a computational model and engine for what I call ‘collective intelligence’, and the Jac programming language to articulate code within this model—so I could strike a balance between the high discovery and nimbleness of AI experimentation in academia and the real-world practicality of the production landscape and scope of sophistication of complex solutions.
Getting AI right is really challenging. When building an end-to-end solution, a single AI model/technique isn’t enough to capture the full experience. Conversational AI is one of the best examples of this.
Why’s that? Conversational AI requires solving multiple distinct problems in tandem to create one experience.
Let’s say you use an AI-based virtual assistant. Every query you make must address multiple disparate challenges in AI:
Each of these challenges requires different models and techniques to solve. You may find yourself wanting to quickly try different things, combine models, feed data into different models, or otherwise change the script.
Main problem: conversational AI is multifaceted and complex, and current software computational models aren’t suited for this.
Here’s an example. You may find yourself committing to certain techniques that seem to work. This may even work for a while. Before you know it, as your solution and codebase increases in size and complexity you become evermore locked into the specific approach and design of the solution, making it evermore challenging to try new things as every small change to the fundamental AI approach becomes harder and harder to integrate.
This style of thinking doesn’t work in conversational AI. Sticking to one approach is the wrong approach when you need multiple techniques and models to succeed, especially when new ideas and techniques in the broader field are rapidly being discovered and developed.
That’s exactly what I observed during my journey building Clinc. It was also these exact reasons that led me to create Jaseci.
Current computational models fall short because of the isolated nature in which they operate. What I’m calling ‘collective intelligence’ is the answer to this problem.
What is collective intelligence? It’s a revolutionary approach to traditional computing models—to how we think about data in computers, to what a function or method even is, to how we address scope (variables) in sophisticated AI engines. Collective intelligence is a new realization for modularity, one where computations can live within the data and being able to articulate situations where data is performing its own computations.
You can think of collective intelligence like an ant colony. In a colony, every ant is free to pursue individual tasks while still working towards the colony’s collective goals. Ants have an instinctive nature to collectivize and work towards the common good while they are perfectly modular and individual neural units unaware of the overarching beauty, complexity, and sophistication of the overarching colony. You can add another ant to do work, while not disrupting the colony, or remove an ant at any time, and the colony continues to operate. Collective intelligence has those same goals. It’s through this collective intelligence that the sophistication of the colony is realized. That is the foundational principle of Jaseci.
But the collective intelligence embodied by Jaseci is nothing without the right programming language to articulate solutions. That’s where Jac comes into the picture.
A key principle of Jaseci is that all data can be represented in various layers of graphs, and all computation can be represented as a robot or ant (walker) walking the edges and nodes of that graph.
Now imagine if that walker was capable of traversing data in its own way with it’s own thoughts. In this modular system, you could have countless walkers doing different things and collaborate later. You could replace a traditional function or method and replace it with a robot you coded. That means replacing arrays, queries, and data with this graph representation of a node/edge. Those same nodes and edges could even have their own in-data computations and data embedded within itself.
That’s exactly what Jaseci actualizes. It’s unlike any language or computational engine that exists today. With this programming language, AI can be fused with the data itself as holistic thinking occurs across disparate AI models and tech while walkers can traverse this data thinking and making small changes as it goes. Indeed I’ve proved formally that all problems in computer science can be solved this way (or this approach is ‘turing complete’).
Because walkers can think and engage with all AI technology and processes, you get a modular way to try things nimbly. You can change the whole way you want to approach reasoning about your data using multiple AI models, flipping the script, and encapsulating a new vision with another walker.
That new robot can traverse the same data (representation of the problem) to think differently. This gives you diversity across your robots.
Think of Jaseci as a new programming language that allows for diversity of thought that allows your computational model to negotiate which solutions are best to determine the best answers that will provide the most value.
This is just the tip of the iceberg. I haven’t even mentioned higher dimensional layers of graphs and how thinking can occur simultaneously across different dimensions of graphs. More on that later. I’m still working on the official Jaseci book.
You may have heard of the Unreal Engine—a popular full-source engine used in the video game industry. The same engine used to power Fortnite, Bioshock, Borderlands 2, and many other games.
My goal with Jaseci Labs is to create the Unreal Engine of AI—a sophisticated platform built for AI solutions.
Think of Jaseci as an enabler for mass-scale production-ready POCs that allows more developers to leverage the full potential of AI and collective intelligence to create innovative new technologies, video games, and other forms of media and entertainment.
This tech stack and programming language are already being used to power several of my own projects. Here are a few examples:
These examples are only a glimpse of what you can do with Jaseci. If you’re interested in learning more about Jaseci, I am open to collaboration.
It’s clear the underlying technology that powers conversational AI needs to change. This is the exact reason why I created Jaseci Labs in the first place.
And anyone that understands my approach to business knows that I’m a show don’t tell kind of guy.
All of the technologies I mentioned above are designed to demonstrate experiences unlike anything you’ve ever seen before because all of them were built on Jaseci.
Sign up to receive updates on the release and be one of the first to trial Zero Shot in the market!