Group items tagged

"In a paper titled "Towards Causal Representation Learning," researchers at the Max Planck Institute for Intelligent Systems, the Montreal Institute for Learning Algorithms (Mila), and Google Research discuss the challenges arising from the lack of causal representations in machine learning models and provide directions for creating artificial intelligence systems that can learn causal representations."

"But ChatGPT and similar programs are, by design, unlimited in what they can "learn" (which is to say, memorize); they are incapable of distinguishing the possible from the impossible. Unlike humans, for example, who are endowed with a universal grammar that limits the languages we can learn to those with a certain kind of almost mathematical elegance, these programs learn humanly possible and humanly impossible languages with equal facility. Whereas humans are limited in the kinds of explanations we can rationally conjecture, machine learning systems can learn both that the earth is flat and that the earth is round. They trade merely in probabilities that change over time."

"This is the broad goal of a new field known as embodied AI, and Li's not the only one embracing it. It overlaps with robotics, since robots can be the physical equivalent of embodied AI agents in the real world, and reinforcement learning - which has always trained an interactive agent to learn using long-term rewards as incentive. But Li and others think embodied AI could power a major shift from machines learning straightforward abilities, like recognizing images, to learning how to perform complex humanlike tasks with multiple steps, such as making an omelet."

"It is in the nature of AI that makers do not, and often cannot, predict what their creations do. We know how to make machines learn. But programmers do not understand completely the knowledge that intelligent computing acquires. If we did, we wouldn't need computers to learn to learn."

"This type of approach can speed up learning times and improve the efficiency of algorithms, says Max Jaderberg at Google's AI company DeepMind. The company used a similar technique last year to teach an AI to explore a virtual maze. Its algorithm learned much more quickly than conventional reinforcement learning approaches. "Our agent is far quicker and requires a lot less experience from the world to train, making it much more data efficient," he says."

"These experiments in computational creativity are enabled by the dramatic advances in deep learning over the past decade. Deep learning has several key advantages for creative pursuits. For starters, it's extremely flexible, and it's relatively easy to train deep-learning systems (which we call models) to take on a wide variety of tasks."

"LeCun argues that even self-supervised learning and learnings from neurobiology won't be enough to achieve artificial general intelligence (AGI), or the hypothetical intelligence of a machine with the capacity to understand or learn from any task. That's because intelligence - even human intelligence - is very specialized, he says. "AGI does not exist - there is no such thing as general intelligence," said LeCun. "We can talk about rat-level intelligence, cat-level intelligence, dog-level intelligence, or human-level intelligence, but not artificial general intelligence.""

""Could teaching be trumped by a learning machine? Are we beginning to glimpse the possibility of machines that teach themselves to teach? They learn what works, what doesn't and deliver ever better performance. We see the embryonic evidence for this in adaptive learning systems, that are truly algorithmic, and do use machine learning, to improve as they deliver. The more students they teach, the better they get. They even tech themselves. This is not science fiction. This is real AI, in real software, delivering real courses, in real institutions." - Donald Clark"

"Whereas most machine learning algorithms can't hone their skills beyond an initial training period, the researchers say the new approach, called a liquid neural network, has a kind of built-in "neuroplasticity." That is, as it goes about its work-say, in the future, maybe driving a car or directing a robot-it can learn from experience and adjust its connections on the fly."

"Brain scans predict students' learning better than exam results and show the underlying structure of thinking. According to recent research published in Science Advances, the conventional exams and grades that schools have long employed may evaluate learning less accurately than brain scans."

"It is one of the first programs to combine an external memory with an approach called deep-learning, in which the program learns how to do tasks independently rather than being pre-programmed with a set of rules by a human."

"Learning from rewards seems like the simplest thing. I make coffee, I sip coffee, I'm happy. My brain registers "brewing coffee" as an action that leads to a reward. That's the guiding insight behind deep reinforcement learning, a family of algorithms that famously smashed most of Atari's gaming catalog and triumphed over humans in strategy games like Go. Here, an AI "agent" explores the game, trying out different actions and registering ones that let it win."

"DeepMind is trying to combine deep learning and algorithms, creating the one algorithm to rule them all: a deep learning model that can learn how to emulate any algorithm, generating an algorithm-equivalent model that can work with real-world data."

"AI tool GNoME finds 2.2 million new crystals, including 380,000 stable materials that could power future technologies Modern technologies from computer chips and batteries to solar panels rely on inorganic crystals. To enable new technologies, crystals must be stable otherwise they can decompose, and behind each new, stable crystal can be months of painstaking experimentation. Today, in a paper published in Nature, we share the discovery of 2.2 million new crystals - equivalent to nearly 800 years' worth of knowledge. We introduce Graph Networks for Materials Exploration (GNoME), our new deep learning tool that dramatically increases the speed and efficiency of discovery by predicting the stability of new materials."

"A core selling point of machine learning is discovery without understanding, which is why errors are particularly common in machine-learning-based science. Three years ago, we compiled evidence revealing that an error called leakage - the machine learning version of teaching to the test - was pervasive, affecting hundreds of papers from 17 disciplines. Since then, we have been trying to understand the problem better and devise solutions.  This post presents an update. In short, we think things will get worse before they get better, although there are glimmers of hope on the horizon."

"Cancer diagnosis and management depend upon the extraction of complex information from microscopy images by pathologists, which requires time-consuming expert interpretation prone to human bias. Supervised deep learning approaches have proven powerful, but are inherently limited by the cost and quality of annotations used for training. Therefore, we present Histomorphological Phenotype Learning, a self-supervised methodology requiring no labels and operating via the automatic discovery of discriminatory features in image tiles. Tiles are grouped into morphologically similar clusters which constitute an atlas of histomorphological phenotypes (HP-Atlas), revealing trajectories from benign to malignant tissue via inflammatory and reactive phenotypes. These clusters have distinct features which can be identified using orthogonal methods, linking histologic, molecular and clinical phenotypes. Applied to lung cancer, we show that they align closely with patient survival, with histopathologically recognised tumor types and growth patterns, and with transcriptomic measures of immunophenotype. These properties are maintained in a multi-cancer study."

"In the short term, generative AI will replace a lot of people because productivity increases while demand stays the same due to inertia. In the long term, the creation of new jobs compensates for the loss of old ones, resulting in a net positive outcome for humans who leave behind jobs no one wants to do. The most important aspect of any technological revolution is the transition from before to after. Timing and location matters: older people have a harder time reinventing themselves into a new trade or craft. Poor people and poor countries have less margin to react to a wave of unemployment. Digital automation is quicker and more aggressive than physical automation because it bypasses logistical constraints-while ChatGPT can be infinitely cloned, a metallic robot cannot. Writing and painting won't die because people care about the human factor first and foremost; there are already a lot of books we can't possibly read in one lifetime so we select them as a function of who's the author. Even if you hate OpenAI and ChatGPT for being responsible for the lack of job postings, I recommend you ally with them for now; learn to use ChatGPT before it's too late to keep your options open. Companies are choosing to reduce costs over increasing output because the sectors where generative AI is useful can't artificially increase demand in parallel to productivity. (Who needs more online content?) Our generation is reasonably angry at generative AI and will bravely fight it. Still, our offspring-and theirs-will be grateful for a transformed world whose painful transformation they didn't have to endure. Certifiable human-made creative output will reduce its quantity but multiply its value in the next years because demand specific for it will grow; automation can mimic 99% of what we do but never reaches 100%. The maxim "AI won't take your job, a person using AI will; yes, you using AI will replace yourself not using it" applies more in the long term than the

"What you may not realize, though, is that machine learning is already all around you, and it can exert a surprising degree of influence over your life. Don't believe me? You might be surprised."

" Critics have pointed out that the old computing adage "garbage in, garbage out" also applies to ML. If the data from which a machine "learns" is biased, then the outputs will reflect those biases. And this could become generalised: we may have created a technology that - however good it is at recommending films you might like - may actually morph into a powerful amplifier of social, economic and cultural inequalities."

"FOLLOW US Twitter / Facebook / RSS Machine learning image classifiers use context clues to help understand the contents of a room, for example, if they manage to identify a dining-room table with a high degree of confidence, that can help resolve ambiguity about other objects nearby, identifying them as chairs. The downside of this powerful approach is that it means machine learning classifiers can be confounded by confusing, out-of-context elements in a scene, as is demonstrated in The Elephant in the Room, a paper from a trio of Toronto-based computer science academics."