Sep 22, 2020

3 Types of Artificial Intelligence

Artificial intelligence
ai systems
Amber Naylor
3 min
Exploring how AI Technology surpasses humans and human intelligence
Exploring how AI Technology surpasses humans and human intelligence...

 

Being the one of the most astounding and complex ‘theory of mind’ creations from humanity, Artificial Intelligence remains largely unexplored, resulting in the opportunity to optimize each AI application that has already been created. Each application not only represents how far we have come, but also how much we have left to develop and explore. Collated from Forbes ’ take on Artificial Intelligence, we outline the main 3 types of AI systems. 

In order to separate ‘types’ of Artificial Intelligence, it first must be understood what criteria they are judged against. Seeming that AI is a system used to replicate human behaviours and capabilities, such as to perform tasks in a way to mimic human abilities, this is the criteria in which the types of Artificial Intelligence are placed. Depending on how the specific intelligent machines compare to the equivalent of a human in terms of performance, it can then be placed classified under one or more types of AI. Under such umbrellas, an AI that is able to perform more human-like functions with equivalent levels of cognitive ability, is noted as a more evolved type. With this in mind, working with this system of classification, there are four types of AI-based systems: Reactive Machines, Limited Memory Machines, Theory of Minds and Self-Aware AI

  1. Reactive Machines

Standing as one of the oldest forms of AI systems, reactive machines have exceedingly limited capabilities. They are able to form the human mind’s ability to respond to various forms of stimuli, but have no memory-based functionality. Due to this, they are unable to use previously gained experiences to aid in their present actions; not having the ability to ‘learn’. In terms of usefulness, the machines could only be used for giving an automatic response to a select set of inputs. A well known example of a reactive AI machine is IBM’s Deep Blue, a machine created to beat chess Grandmaster, Garry Kasparov in 1997.

2. Theory of Mind

The Theory of Mind type of AI is currently still being developed and researched. It is the next level of Artificial Intelligence systems that have been created from researchers, still exploring ways to innovate. The difference between This new form of AI system is the way that the machine will be able to interact. It will be able to discern between and entities needs, beliefs, emotions and thought processes. Although Artificial Emotional Intelligence already has its own holding with the industry, the Theory of Mind branches out so much further. With the demand for understanding human needs, the machines will have the ability to perceive humans as individuals, knowing that their minds can be changed and focused on many areas; essentially, ‘understanding’ a human. 

3. Self-aware

Self-aware AI currently only exists hypothetically. Still under development, it will be an Artificial Intelligence system that has been evolved to the next level; so akin to the human brain that it is able to find its own form of self-awareness. To be able to develop this type of AI, which is still decades - if not centuries away, will always be the ultimate end goal of all AI research. This type of machine will not only be able to interpret the emotions of the individual it is interacting with, but also develop emotions, needs and potentially desires of its own. There are individuals who believe that the development of the self-aware will not have the enlightened results as hoped for. Although there are obvious benefits for civilization with the progression of this type of AI, some consider it as a lead to catastrophe. If AI machines were truly self-aware, they would be able to consider feelings such as self-preservation, which may indirectly or directly affect humanity and the way we live our day-to-day lives. 

Share article

Jun 17, 2021

Facebook Develops AI to Crackdown on Deepfakes

Facebook
MSU
AI
Deepfakes
3 min
Social media giant, Facebook, has developed artificial intelligence that can supposedly identify and reverse-engineer deepfake images

In light of the large tidal wave of increasingly believable deepfake images and videos that have been hitting the feeds of every major social media and news outlet in recent years, global organisations have started to consider the risk factor behind them. While the majority of deepfakes are created purely for amusement, their increasing sophistication is leading to a very simple question: What happens when a deepfake is produced not for amusement, but for malicious intent on a grander scale? 

 

Yesterday, Facebook revealed that it was also concerned by that very question and that it had decided to take a stand against deepfakes. In partnership with Michigan State University, the social media giant presented “a research method of detecting and attributing deepfakes that relies on reverse engineering from a single AI-generated image to the generative model used to produce it.” 

 

The promise is that Facebook’s method will facilitate deepfake detection and tracing in real-world settings, where the deepfake image itself is often the only information detectors have to work with. 

Why Reverse Engineering? 

Right now, researchers identify deepfakes through two primary methods: detection, which distinguishes between real and deepfake images, and image attribution, which identifies whether the image was generated using one of the AI’s training models. But generative photo techniques have advanced in scale and sophistication over the past few years, and the old strategies are no longer sufficient. 

 

First, there are only so many images presented in AI training. If the deepfake was generated by an unknown, alternative model, even artificial intelligence won’t be able to spot it—at least, until now. Reverse engineering, common practice in machine learning (ML), can uncover unique patterns left by the generating model, regardless of whether it was included in the AI’s training set. This helps discover coordinated deepfake attacks or other instances in which multiple deepfakes come from the same source. 

 

How It Works 

Before we could use deep learning to generate images, criminals and other ill-intentioned actors had a limited amount of options. Cameras only had so many tools at their disposal, and most researchers could easily identify certain makes and models. But deep learning has ushered in an age of endless options, and as a result, it’s grown increasingly difficult to identify deepfakes.

 

To counteract this, Facebook ran deepfakes through a fingerprint estimation network (FEN) to estimate some of their details. Fingerprints are essentially patterns left on an image due to manufacturing imperfections, and they help identify where the image came from. By evaluating the fingerprint magnitude, repetition frequency, and symmetrical frequency, Facebook then applied those constraints to predict the model’s hyperparameters. 

 

What are hyperparameters? If you imagine a generative model as a car, hyperparameters are similar to the engine components: certain properties that distinguish your fancy automobile from others on the market. ‘Our reverse engineering technique is somewhat like recognising [the engine] components of a car based on how it sounds’, Facebook explained, ‘even if this is a new car we’ve never heard of before’. 

 

What Did They Find? 

‘On standard benchmarks, we get state-of-the-art results’, said Facebook research lead Tal Hassner. Facebook added that the fingerprint estimation network (FEN) method can be used for not only model parsing, but detection and image attribution. While this research is the first of its kind, making it difficult to assess the results, the future looks promising. 


Facebook’s AI will introduce model parsing for real-world applications, increasing our understanding of deepfake detection. As cybersecurity attacks proliferate, and generative AI falls into the hands of those who would do us harm, this method could help the ‘good guys’ stay one step ahead. As Hassner explained: ‘This is a cat-and-mouse game, and it continues to be a cat-and-mouse game’.

Share article