{"id":6318,"date":"2021-07-05T15:14:00","date_gmt":"2021-07-05T15:14:00","guid":{"rendered":"http:\/\/TheNextWeb=1359302"},"modified":"2021-07-05T15:14:00","modified_gmt":"2021-07-05T15:14:00","slug":"pioneers-of-deep-learning-ai-think-its-future-is-gonna-be-lit","status":"publish","type":"post","link":"https:\/\/www.londonchiropracter.com\/?p=6318","title":{"rendered":"Pioneers of deep learning AI think its future is gonna be lit"},"content":{"rendered":"\n

Deep neural networks will move past their shortcomings without help from <\/span>symbolic artificial intelligence<\/a>, three pioneers of deep learning argue in a paper published in the July issue of the <\/span>Communications of the ACM <\/span><\/em>journal.<\/p>\n

In their paper, Yoshua Bengio, Geoffrey Hinton, and Yann LeCun, recipients of the 2018 Turing Award, explain the current <\/span>challenges of deep learning<\/a> <\/span>and how it differs from learning in humans and animals. They also explore recent advances in the field that might provide blueprints for the future directions for research in deep learning.<\/p>\n

Titled \u201cDeep Learning for AI<\/a>,\u201d the paper envisions a future in which deep learning models can learn with little or no help from humans, are flexible to changes in their environment, and can solve a wide range of reflexive and cognitive problems.<\/p>\n

The challenges of deep learning<\/h2>\n

Deep learning<\/a> <\/span>is often compared to the brains of humans and animals. However, the past years have proven that artificial neural networks, the main component used in deep learning models, lack the efficiency, flexibility, and versatility of their biological counterparts.<\/p>\n

In their paper, Bengio, Hinton, and LeCun acknowledge these shortcomings. \u201cSupervised learning, while successful in a wide variety of tasks, typically requires a large amount of human-labeled data. Similarly, when reinforcement learning is based only on rewards, it requires a very large number of interactions,\u201d they write.<\/p>\n

Supervised learning<\/a> <\/span>is a popular subset of machine learning algorithms, in which a model is presented with labeled examples, such as a list of images and their corresponding content. The model is trained to find recurring patterns in examples that have similar labels. It then uses the learned patterns to associate new examples with the right labels. Supervised learning is especially useful for problems where labeled examples are abundantly available.<\/p>\n

Reinforcement learning<\/a> <\/span>is another branch of machine learning, in which an \u201cagent\u201d learns to maximize \u201crewards\u201d in an environment. An environment can be as simple as a tic-tac-toe board in which an AI player is rewarded for lining up three Xs or Os, or as complex as an urban setting in which a <\/span>self-driving car<\/a> <\/span>is rewarded for avoiding collisions, obeying traffic rules, and reaching its destination. The agent starts by taking random actions. As it receives feedback from its environment, it finds sequences of actions that provide better rewards.<\/p>\n

In both cases, as the scientists acknowledge, machine learning models require huge labor. Labeled datasets are hard to come by, especially in specialized fields that don\u2019t have public, open-source datasets, which means they need the hard and expensive labor of human annotators. And complicated reinforcement learning models require massive computational resources to run a vast number of training episodes, which makes them available to a few, <\/span>very wealthy AI labs and tech companies<\/a>.<\/p>\n

Bengio, Hinton, and LeCun also acknowledge that current deep learning systems are still <\/span>limited in the scope of problems they can solve<\/a>. They perform well on specialized tasks but \u201care often brittle outside of the narrow domain they have been trained on.\u201d Often, slight changes such as <\/span>a few modified pixels<\/a> <\/span>in an image or a very slight alteration of rules in the environment can cause deep learning systems to go astray.<\/p>\n

The brittleness of deep learning systems is largely due to machine learning models being based on the \u201cindependent and identically distributed\u201d (i.i.d.) assumption, which supposes that real-world data has the same distribution as the training data. i.i.d also assumes that observations do not affect each other (e.g., coin or die tosses are independent of each other).<\/p>\n

\u201cFrom the early days, theoreticians of machine learning have focused on the iid assumption\u2026 Unfortunately, this is not a realistic assumption in the real world,\u201d the scientists write.<\/p>\n

Real-world settings are constantly changing due to different factors, many of which are virtually impossible to represent <\/span>without causal models<\/a>. Intelligent agents must constantly observe and learn from their environment and other agents, and they must adapt their behavior to changes.<\/p>\n

\u201c[T]he performance of today\u2019s best AI systems tends to take a hit when they go from the lab to the field,\u201d the scientists write.<\/p>\n

The i.i.d. assumption becomes even more fragile when applied to fields such as <\/span>computer vision<\/a> <\/span>and natural language processing, where the agent must deal with high-entropy environments. Currently, many researchers and companies try to overcome the limits of deep learning by <\/span>training neural networks on more data<\/a>, hoping that larger datasets will cover a wider distribution and reduce the chances of failure in the real world.<\/p>\n

Deep learning vs hybrid AI<\/h2>\n