intelligence artificielle generalisee deepmind gato

Google aurait créé la première intelligence artificielle généralisée, “concurrençant” l’esprit humain

⇧ [VIDÉO] You might also like this partner content (after ad)

DeepMind, a company (belonging to Google) specialized in artificial intelligence, has just presented its new artificial intelligence named “Gato”. Unlike “classic” AIs, which specialize in a specific task, Gato is able to perform more than 600 tasks, often much better than humans. Controversy is launched as to whether this is really the first “generalized artificial intelligence” (GAI). Experts remain skeptical of DeepMind’s announcement.

Artificial intelligence has positively changed many disciplines. Incredible specialized neural networks are now able to produce results far beyond human capabilities in many areas.

One of the great challenges in the field of AI is the realization of a system integrating generalized artificial intelligence (GAI), or strong artificial intelligence. Such a system must be able to understand and master any task of which a human being would be capable. She would therefore be able to compete with human intelligence, and even develop a certain degree of consciousness. Earlier this year, Google unveiled a type of AI capable of coding like an average programmer. Recently, in this race for AI, DeepMind announced the creation of Gato, an artificial intelligence presented as the first AGI in the world. The results are published in arXiv.

An unprecedented generalist agent model

A single AI system capable of solving many tasks is not something new. For example, Google recently started using a system for its search engine called the “unified multitasking model,” or MUM, that can handle text, images, and video to perform tasks, from research to cross-linguistic variations. in the spelling of a word, and the association of search queries with relevant images.

Incidentally, Senior Vice President Prabhakar Raghavan provided an impressive example of MUM in action, using the mock search query: I hiked Mount Adams and now want to hike Mount Fuji next fall, what should I do differently to prepare? “. MUM enabled Google Search to show the differences and similarities between Mount Adams and Mount Fuji. He also brought up articles dealing with the equipment needed to ascend the latter. Nothing too impressive you would say, but concretely with Gato, what is innovative is the diversity of the tasks that are approached and the method of training, of a single and unique system.

Gato’s guiding design principle is to train on the widest variety of relevant data possible, including various applications such as images, text, proprioception, joint torques, button presses and others. discrete and continuous observations and actions.

To enable processing of this multimodal data, scientists encode it into a flat sequence of “tokens”. These tokens are used to represent data in a way that Gato can understand, allowing the system, for example, to figure out which combination of words in a sentence makes grammatical sense. These sequences are grouped together and processed by a transformative neural network, typically used in language processing. The same network, with the same weights, is used for the different tasks, unlike traditional neural networks. Indeed, in the latter, each neuron is assigned a particular weight and therefore a different importance. In simple terms, the weight determines what information enters the network and calculates an output data.

In this representation, Gato can be trained and sampled from a standard large-scale language model, on a large number of datasets including agents’ experience in simulated and real environments, in addition to a variety of natural language datasets and images. In its operation, Gato uses context to assemble these sampled tokens to determine the form and content of its responses.

Example of execution of Gato. The system “consumes” a sequence of previously sampled observation and action tokens to produce the next action. The new action is applied, by the agent (Gato), to the environment (a game console in this illustration), a new set of observations is obtained, and the process repeats. © S. Reed et al., 2022.

The results are quite heterogeneous. When it comes to dialog, Gato falls far short of rivaling the prowess of GPT-3, Open AI’s text generation model. He can give wrong answers during conversations. For example he answers that Marseille is the capital of France. The authors point out that this could probably be improved with further scaling.

Nevertheless, he still proved to be extremely capable in other areas. Its designers claim that, half the time, Gato performs better than human experts in 450 of the 604 tasks listed in the research paper.

deepmind gato token sequence
Examples of the tasks performed by Gato, as sequences of tokens. © S. Reed et al., 2022.

The Game is Over “, Actually ?

Some AI researchers see the AGI as an existential catastrophe for humans: a “super intelligent” system that surpasses human intelligence would replace humanity on Earth, under the worst-case scenario. Other experts believe that it will not be possible in our lifetime to see the emergence of these AGIs. This is the pessimistic opinion that Tristan Greene argued in his editorial on the site TheNextWeb. He explains that it’s easy to mistake Gato for a real IAG. The difference, however, is that a general intelligence could learn to do new things without prior training.

The response to this article was not long in coming. On TwitterNando de Freitas, researcher at DeepMind and professor of machine learning at the University of Oxford, said the game was over (“ The Game is Over ”) in the long quest for generalized artificial intelligence. He adds : ” It’s about making these models bigger, safer, more computationally efficient, faster to sample, with smarter memory, more modalities, innovative data, online/offline… It is by solving these challenges that we will obtain the IAG “.

Nevertheless, the authors warn against the development of these AGIs: “ Although generalist agents are still an emerging field of research, their potential impact on society calls for a thorough interdisciplinary analysis of their risks and benefits. […] Generalist agent harm mitigation tools are relatively underdeveloped and require further research before these agents are deployed “.

Moreover, generalist agents, capable of performing actions in the physical world, pose new challenges requiring new mitigation strategies. For example, physical embodiment could lead users to anthropomorphize the agent, leading to misplaced trust in the case of a faulty system.

In addition to these risks of seeing the AGI tip into a harmful operation for humanity, no data currently demonstrates the ability to produce solid results in a consistent manner. This is particularly due to the fact that human problems are often difficult, not always having a single solution, and for which no prior training is possible.

Tristant Greene, despite Nando de Fraitas’ response, maintains his opinion just as harshly, on TheNextWeb : “ It’s nothing short of miraculous to see a machine pull off feats of diversion and conjuring a la Copperfield, especially when you realize that said machine is no smarter than a toaster (and obviously dumber than the dumbest mouse) “.

Whether or not we agree with these statements, or whether we are more optimistic about the development of AGIs, it nevertheless seems that the scaling up of such intelligences, competing with our human minds, is still far from complete. be completed, and the controversies appeased.

Source: arXiv

#Google #aurait #créé #première #intelligence #artificielle #généralisée #concurrençant #lesprit #humain

This will close in 0 seconds