With their great deterministic and scalable structure, neural network architectures can accumulate a lot of data in an exponential rate.
Yet, what makes them learning so fast is probably what prevents them from a deeper understanding of their environment.
In fact, every part of their structure is already determined by their function, and thus leaves no room for organic complexity. In contrast, each description of the human brain cannot predict the resulting behavior.
Michael S. Gazzaniga, in “the Consciousness Instinct”, brings this back to the semantic and arbitrary part of the DNA: every time RNA translates DNA there is always…
Through academic contests these past 20 years(Imagenet, Alphago challenge, Word2Vec…), AI has shown huge successes in specific fields.
Whether object, voice or word recognition, deep learning algorithms from major tech companies have shown accuracy close to perfection.
But these results hide imperfections that could soon hinder their progress. According to Melanie Mitchell in A Guide for thinking humans, AI lacks the basis of a generalist intelligence that understands and adapts to situations, and that could significantly slow down data learning.
Here are four cognitive skills that machines lack, and the skills they need to pass these obstacles.
As deep learning technologies are growing, AI has proven to be a real success in fields like visual or voice recognition. Algorithms have refined their ability to make sense of image and speech based on statistical regularities. However, these analytical skills are increasingly showing their limitations in gaining a deeper understanding of their objects.
These computers cannot really understand the meaning of a sentence or an image by relying like humans on disparate data sources. They are unable to connect and use knowledge from different fields that have no apparent similarity.
Computing science therefore needs algorithms that can think in…
Using a machine that obeys every command with extreme redundancy, coders easily feel like the isolated creator of new worlds. And with the exponential growth of AI technologies, they can lock themselves even more in their work, convinced by the necessity of their algorithm-centered vision.
But this image can be to the detriment of both programmers and users. Coders have a vested interest in considering the impact of their technologies on humans, and in reflecting on the ethics and values of their creations.
Here’s how they can consider their work more in this perspective, and thus make it more meaningful.
Promoted as a brand new business model, the gig economy is receiving more and more criticisms. Instead of giving new freedom to workers, gig platforms lock a large majority of them into highly precarious situations.
According to Jeremias Prassl in Humans as a service, these platforms use AI technologies to exert more control on human workers, rather than to increase their trading power and opportunities. As a result, they keep for themselves the productive and economic benefits of IA.
Here’s how they could put AI technologies at the service of the elevation of human skills.
Many economists have wondered about the economic stagnation that has taken place since the 21st century. While the information technology was booming, the GDP growth has unexpectedly gone from an average of 2–3% to 0–1%.
For some, this slowdown is due to rising inequalities, growing job automation, or demographic maturation.
But according to Dietrich Vollrath in Fully Grown, one of the key explanations often overlooked, is the transition from an industrial economy to a service economy, where productivity gains are much lower.
It is difficult to replace the value brought by 1 hour of consultation with a doctor, a nurse…
Since the ancient invention of the wooden leg, designers have made great strides in technologies that push the boundaries of human capabilities. In particular, neuroscientists have understood how resilient the human brain is and how it can capture very different types of sensory and motor data.
In fact, according to David Eagleman in LiveWired, the brain can read any sensory information (auditory, visual, tactile) regardless of their various origins, as long as they translate as electrical stimuli.
This makes possible technologies that allow us to perceive non-human signals (such as magnetic field or new light spectrum) and to manipulate objects…
Before being a danger to others, men are a danger to themselves. As white men from western societies, they are 3 times more likely to experience episodes of depression, mental complications and serious health problems.
Men are sick of themselves, but it is not really their fault. Having been educated to an independent and Stoic masculine ideal, they have distanced themselves from their emotions and care, and refrain from a masculine identity that they truly desire.
Yet, they have no interest in victimizing themselves. On the contrary, they have every interest in engaging in the same struggle as women, namely…
As machine learning is collecting data at an exponential rate, AI isn’t only trying to copy human-learn behaviors. They are also showing vibrant creativity in fields such as games, visual and musical arts. Even if they are still limited, some neural-trained algorithms can already assess the strategies of great artists and masters and create confusingly human-like artworks.
Without questioning human skills, these robotic capabilities seem to complement our human intuition and inspire creators to go deeper in their creative process and ideas.
Here is the example of 4 robots remarkable by their creativity, and what they teach to human creators.
Since I learned SEO in a web agency, I’ve always been sidelined between two common strategies. Either optimize my content for end-users through genuine and human-made content. Or optimize it for Google through technical shortcuts and generated content.
For Google, the debate has always been straightforward: human input is everything, so according to their guidelines you better create by yourself reliable, accurate and relevant content.
Yet, as machine learning technologies are growing, generated content might be proliferating on the Internet, without having a way to differentiate it from human content. …