Down the Rabbit Hole: July 16th, 2021
Start your day with the latest advancements in analytics, data science and automation…
DeepMind scientists: Reinforcement learning is enough for general AI
In a new paper submitted to the peer-reviewed Artificial Intelligence journal, scientists at UK-based AI lab DeepMind argue that intelligence and its associated abilities will emerge not from formulating and solving complicated problems but by sticking to a simple but powerful principle: reward maximization.
Titled “Reward is Enough,” the paper draws inspiration from studying the evolution of natural intelligence as well as drawing lessons from recent achievements in artificial intelligence.
The authors conclude that reinforcement learning, a branch of AI that is based on reward maximization, can lead to the development of artificial general intelligence.
How businesses can safeguard against rogue AI
Three decades after a US university student called Robert Tappan Morris was convicted of launching the first widely known malware attack on the internet, cybercrime has become big business, costing the global economy an estimated £2.1m a minute.
Facing this relentless onslaught, organisations of all sizes have had to up their game. For many, artificial intelligence has become key, as it can detect irregularities and pinpoint potential threats with much greater accuracy than traditional defences.
Honey I Shrunk the Model: Why Big Machine Learning Models Must Go Small
Research from OpenAI showed that between 2012 and 2018, computing power for deep learning models grew a shocking 300,000x, outpacing Moore’s Law.
The problem lies not only in training these algorithms, but also running them in production, or the inference phase. For many teams, practical use of deep learning models remains out of reach, due to sheer cost and resource constraints.
Pruning, quantization, and transfer learning are three specific techniques that could democratize machine learning for organizations who don’t have millions of dollars to invest in moving models to production. This is especially important for “edge” use cases, where larger, specialized AI hardware is physically impractical.
