AI News, 194 Best Artificial Intelligence Technology artificial intelligence

194. CRISPR Convergence

[Editor’s Note: In today’s post, returning guest blogger and proclaimed Mad Scientist Howard R.

Comparing the genetic engineering tool Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) to the internet in terms of its revolutionary potential, Mr. Simkin examines three scenarios where this capability could be harnessed for nefarious purposes.

It’s just not very evenly distributed.” – William Gibson, science fiction author who coined the word cyberspace in 1984.1 Purpose: This paper briefly examines the convergence of trends in technology as they affect CRISPR2 technology through the lens of three possible users of the technology – the Democratic People’s Republic of Korea (DPRK), a future Aum Shinrikyo3 -like entity, and a Unabomber-like4 super-empowered individual.

survey of the two most commonly available, authoritative sources on the FOE points to an ever-increasing rate of technological change, the growth of mega-cities, and the diffusion of cutting-edge technology into the hands of both state and non-state actors as well as super-empowered individuals.5 Over the next ten to twenty years, the world will experience dramatic changes in technology.

Governments and businesses are investing billions of dollars into research programs and tech startups associated with all five of the technological fields shown in Figure 1.6 The convergence of these technologies, impelled by increasingly capable Artificial Intelligence (AI) will drive change that will approximate that of Moore’s Law – doubling in power while halving in cost every two years.

Our adversaries – states, non-state actors, and super-empowered individuals – will undoubtedly seek to harness these trends to accomplish their ends.

However, its applications to gene editing have already become significant.7 As the web magazine Futurism observed, “As the accuracy, efficiency, and cost-effectiveness of the system became more and more apparent, researchers and pharmaceutical companies jumped on the technique, modifying it, improving it, and testing it on different genetic issues.”8 This tool could lead to gene editing techniques that could strengthen disease resistance and improve strength and mental abilities.

What was formerly only available at the cost of billions of dollars and years of research can now be achieved by a single individual at a nominal cost.

The original human genome project took ten years, employed a large research team with state-of-the-art laboratories, and cost a billion dollars.

To the point, in 2017 Canadian researchers at the University of Alberta revived an extinct horsepox virus using synthetic DNA strands ordered for about $100,000.

His effort opened up new possibilities for researchers looking to make better vaccines, but also those looking to use these viruses as bioweapons9 including smallpox.10 Questions This causes a number of questions to spring to mind.

The sort of enemy who would employ CRISPR to design bioweapons fits one of three profiles, each of which has their own present day or historical example.

With such underlying beliefs, the end justifies the means when dealing with inferiors.11 It doesn’t take much imagination to see that the DPRK would have no moral or ethical problems with creating an asymptomatic, race-specific, highly contagious and deadly disease.

The original organization employed Sarin in the Tokyo subway in 1995 but it also conducted extensive research and testing into bioweapons to include anthrax, botulinum toxin, and the Ebola virus in 1992 –

of human germline editing in her recent editorial entitled CRISPR’s unwanted anniversary in the journal Science:  “Consequences for defying established restrictions should include, at a minimum, loss of funding and publication privileges.

The future may require the power of AI, data science, big data, and quantum computers to identify and track potential threats.

Cuiker and Mayer-Schoenberg observe that, “Using big data will sometimes mean forgoing the quest for why in return for knowing what.”14 In other words, it involves a shift from understanding causation to seeking a correlation derived from big data to provide a proxy for what you are trying to understand.

Google ran over 50 million search terms through 450 million algorithms before arriving at a list of 45 search terms that –

Using this approach, Google was able to detect warning signs within one or two days of an outbreak, pinpoint the geographic area, and even estimate the percentage of the population affected.

Some components of the solution – like a robust public health system – are already in place in the U.S. The future public health system will rapidly identify the bioweapon and begin to develop treatments.

The government will enforce such measures as social distance, allowing virulent strains to ‘burn out.’ In the future, the scientific community will use AI and quantum computing to run simulations that come up with novel approaches to mitigating the effects of any bioweapon.

In the hands of a rogue nation, a terrorist organization, or a super-empowered individual, it could unleash old diseases such as smallpox or new diseases with no known treatment.

His subject matter expertise includes analyzing and evaluating historical, current and emerging technology as well as Combined, Joint, Multi-Service, Army and ARSOF organizational initiatives, trends, and concepts to determine the implications for ARSOF units.

Deep learning

Deep learning (also known as deep structured learning or hierarchical learning) is part of a broader family of machine learning methods based on artificial neural networks.

Deep learning architectures such as deep neural networks, deep belief networks, recurrent neural networks and convolutional neural networks have been applied to fields including computer vision, speech recognition, natural language processing, audio recognition, social network filtering, machine translation, bioinformatics, drug design, medical image analysis, material inspection and board game programs, where they have produced results comparable to and in some cases superior to human experts.[4][5][6]

Deep learning is a class of machine learning algorithms that[11](pp199–200) uses multiple layers to progressively extract higher level features from the raw input.

For example, in image processing, lower layers may identify edges, while higher layers may identify the concepts relevant to a human such as digits or letters or faces.

Most modern deep learning models are based on artificial neural networks, specifically, Convolutional Neural Networks (CNN)s, although they can also include propositional formulas or latent variables organized layer-wise in deep generative models such as the nodes in deep belief networks and deep Boltzmann machines.[12]

No universally agreed upon threshold of depth divides shallow learning from deep learning, but most researchers agree that deep learning involves CAP depth higher than 2.

For supervised learning tasks, deep learning methods eliminate feature engineering, by translating the data into compact intermediate representations akin to principal components, and derive layered structures that remove redundancy in representation.

The classic universal approximation theorem concerns the capacity of feedforward neural networks with a single hidden layer of finite size to approximate continuous functions.[18][19][20][21][22]

The universal approximation theorem for deep neural networks concerns the capacity of networks with bounded width but the depth is allowed to grow.

proved that if the width of a deep neural network with ReLU activation is strictly larger than the input dimension, then the network can approximate any Lebesgue integrable function;

A 1971 paper described already a deep network with 8 layers trained by the group method of data handling algorithm.[33]

By 1991 such systems were used for recognizing isolated 2-D hand-written digits, while recognizing 3-D objects was done by matching 2-D images with a handcrafted 3-D object model.

But while Neocognitron required a human programmer to hand-merge features, Cresceptron learned an open number of features in each layer without supervision, where each feature is represented by a convolution kernel.

In 1994, André de Carvalho, together with Mike Fairhurst and David Bisset, published experimental results of a multi-layer boolean neural network, also known as a weightless neural network, composed of a 3-layers self-organising feature extraction neural network module (SOFT) followed by a multi-layer classification neural network module (GSN), which were independently trained.

In 1995, Brendan Frey demonstrated that it was possible to train (over two days) a network containing six fully connected layers and several hundred hidden units using the wake-sleep algorithm, co-developed with Peter Dayan and Hinton.[44]

Simpler models that use task-specific handcrafted features such as Gabor filters and support vector machines (SVMs) were a popular choice in the 1990s and 2000s, because of artificial neural network's (ANN) computational cost and a lack of understanding of how the brain wires its biological networks.

These methods never outperformed non-uniform internal-handcrafting Gaussian mixture model/Hidden Markov model (GMM-HMM) technology based on generative models of speech trained discriminatively.[50]

The speaker recognition team led by Larry Heck achieved the first significant success with deep neural networks in speech processing in the 1998 National Institute of Standards and Technology Speaker Recognition evaluation.[53]

principle of elevating 'raw' features over hand-crafted optimization was first explored successfully in the architecture of deep autoencoder on the 'raw' spectrogram or linear filter-bank features in the late 1990s,[53]

Many aspects of speech recognition were taken over by a deep learning method called long short-term memory (LSTM), a recurrent neural network published by Hochreiter and Schmidhuber in 1997.[55]

showed how a many-layered feedforward neural network could be effectively pre-trained one layer at a time, treating each layer in turn as an unsupervised restricted Boltzmann machine, then fine-tuning it using supervised backpropagation.[63]

The impact of deep learning in industry began in the early 2000s, when CNNs already processed an estimated 10% to 20% of all the checks written in the US, according to Yann LeCun.[72]

was motivated by the limitations of deep generative models of speech, and the possibility that given more capable hardware and large-scale data sets that deep neural nets (DNN) might become practical.

However, it was discovered that replacing pre-training with large amounts of training data for straightforward backpropagation when using DNNs with large, context-dependent output layers produced error rates dramatically lower than then-state-of-the-art Gaussian mixture model (GMM)/Hidden Markov Model (HMM) and also than more-advanced generative model-based systems.[64][75]

offering technical insights into how to integrate deep learning into the existing highly efficient, run-time speech decoding system deployed by all major speech recognition systems.[11][77][78]

In 2010, researchers extended deep learning from TIMIT to large vocabulary speech recognition, by adopting large output layers of the DNN based on context-dependent HMM states constructed by decision trees.[80][81][82][77]

In 2009, Nvidia was involved in what was called the “big bang” of deep learning, “as deep-learning neural networks were trained with Nvidia graphics processing units (GPUs).”[83]

In 2014, Hochreiter's group used deep learning to detect off-target and toxic effects of environmental chemicals in nutrients, household products and drugs and won the 'Tox21 Data Challenge' of NIH, FDA and NCATS.[93][94][95]

Although CNNs trained by backpropagation had been around for decades, and GPU implementations of NNs for years, including CNNs, fast implementations of CNNs with max-pooling on GPUs in the style of Ciresan and colleagues were needed to progress on computer vision.[85][87][39][96][2]

In 2013 and 2014, the error rate on the ImageNet task using deep learning was further reduced, following a similar trend in large-scale speech recognition.

For example, in image recognition, they might learn to identify images that contain cats by analyzing example images that have been manually labeled as 'cat' or 'no cat' and using the analytic results to identify cats in other images.

Over time, attention focused on matching specific mental abilities, leading to deviations from biology such as backpropagation, or passing information in the reverse direction and adjusting the network to reflect that information.

Neural networks have been used on a variety of tasks, including computer vision, speech recognition, machine translation, social network filtering, playing board and video games and medical diagnosis.

Despite this number being several order of magnitude less than the number of neurons on a human brain, these networks can perform many tasks at a level beyond that of humans (e.g., recognizing faces, playing 'Go'[105]

The extra layers enable composition of features from lower layers, potentially modeling complex data with fewer units than a similarly performing shallow network.[12]

The training process can be guaranteed to converge in one step with a new batch of data, and the computational complexity of the training algorithm is linear with respect to the number of neurons involved.[121][122]

that involve multi-second intervals containing speech events separated by thousands of discrete time steps, where one time step corresponds to about 10 ms.

All major commercial speech recognition systems (e.g., Microsoft Cortana, Xbox, Skype Translator, Amazon Alexa, Google Now, Apple Siri, Baidu and iFlyTek voice search, and a range of Nuance speech products, etc.) are based on deep learning.[11][128][129][130]

DNNs have proven themselves capable, for example, of a) identifying the style period of a given painting, b) Neural Style Transfer - capturing the style of a given artwork and applying it in a visually pleasing manner to an arbitrary photograph or video, and c) generating striking imagery based on random visual input fields.[134][135]

Word embedding, such as word2vec, can be thought of as a representational layer in a deep learning architecture that transforms an atomic word into a positional representation of the word relative to other words in the dataset;

Deep learning has been shown to produce competitive results in medical application such as cancer cell classification, lesion detection, organ segmentation and image enhancement[170][171]

Finding the appropriate mobile audience for mobile advertising is always challenging, since many data points must be considered and assimilated before a target segment can be created and used in ad serving by any ad server.[172]

'Deep anti-money laundering detection system can spot and recognize relationships and similarities between data and, further down the road, learn to detect anomalies or classify and predict specific events'.

Deep learning is closely related to a class of theories of brain development (specifically, neocortical development) proposed by cognitive neuroscientists in the early 1990s.[177][178][179][180]

These developmental models share the property that various proposed learning dynamics in the brain (e.g., a wave of nerve growth factor) support the self-organization somewhat analogous to the neural networks utilized in deep learning models.

Like the neocortex, neural networks employ a hierarchy of layered filters in which each layer considers information from a prior layer (or the operating environment), and then passes its output (and possibly the original input), to other layers.

Other researchers have argued that unsupervised forms of deep learning, such as those based on hierarchical generative models and deep belief networks, may be closer to biological reality.[184][185]

researchers at The University of Texas at Austin (UT) developed a machine learning framework called Training an Agent Manually via Evaluative Reinforcement, or TAMER, which proposed new methods for robots or computer programs to learn how to perform tasks by interacting with a human instructor.[176]

Such techniques lack ways of representing causal relationships (...) have no obvious ways of performing logical inferences, and they are also still a long way from integrating abstract knowledge, such as information about what objects are, what they are for, and how they are typically used.

systems, like Watson (...) use techniques like deep learning as just one element in a very complicated ensemble of techniques, ranging from the statistical technique of Bayesian inference to deductive reasoning.'[203]

As an alternative to this emphasis on the limits of deep learning, one author speculated that it might be possible to train a machine vision stack to perform the sophisticated task of discriminating between 'old master' and amateur figure drawings, and hypothesized that such a sensitivity might represent the rudiments of a non-trivial machine empathy.[204]

In further reference to the idea that artistic sensitivity might inhere within relatively low levels of the cognitive hierarchy, a published series of graphic representations of the internal states of deep (20-30 layers) neural networks attempting to discern within essentially random data the images on which they were trained[206]

Goertzel hypothesized that these behaviors are due to limitations in their internal representations and that these limitations would inhibit integration into heterogeneous multi-component artificial general intelligence (AGI) architectures.[208]

Learning a grammar (visual or linguistic) from training data would be equivalent to restricting the system to commonsense reasoning that operates on concepts in terms of grammatical production rules and is a basic goal of both human language acquisition[212]

In 2016 researchers used one ANN to doctor images in trial and error fashion, identify another's focal points and thereby generate images that deceived it.

Another group showed that certain psychedelic spectacles could fool a facial recognition system into thinking ordinary people were celebrities, potentially allowing one person to impersonate another.

ANNs can however be further trained to detect attempts at deception, potentially leading attackers and defenders into an arms race similar to the kind that already defines the malware defense industry.

ANNs have been trained to defeat ANN-based anti-malware software by repeatedly attacking a defense with malware that was continually altered by a genetic algorithm until it tricked the anti-malware while retaining its ability to damage the target.[216]

The philosopher Rainer Mühlhoff distinguishes five types of 'machinic capture' of human microwork to generate training data: (1) gamification (the embedding of annotation or computation tasks in the flow of a game), (2) 'trapping and tracking' (e.g.

As Mühlhoff argues, involvement of human users to generate training and verification data is so typical for most commercial end-user applications of Deep Learning that such systems may be referred to as 'human-aided artificial intelligence'[218].

19 Artificial Intelligence Technologies To Look For In 2019

Tech decision makers are (and should keep) looking for ways to effectively implement artificial intelligence technologies into their businesses and, therefore, drive value.

By providing algorithms, APIs (application programming interface), development and training tools, big data, applications and other machines, ML platforms are gaining more and more traction every day.

The last one is actually the first and only audience management tool in the world that applies real AI and machine learning to digital advertising to find the most profitable audience or demographic group for any ad.

Deep learning platforms use a unique form of ML that involves artificial neural circuits with various abstraction layers that can mimic the human brain, processing data and creating patterns for decision making.

It allows for more natural interactions between humans and machines, including interactions related to touch, image, speech, and body language recognition, and is big within the market research field.

Robotic processes automation uses scripts and methods that mimic and automate human tasks to support corporate processes.  It is particularly useful for situations when hiring humans for a specific job or task is too expensive or inefficient.

Their digital twins are basically, lines of software code, but the most elaborate versions look like 3-D computer-aided design drawings full of interactive charts, diagrams, and data points.

AI and ML are now being used to move cyberdefense into a new evolutionary phase in response to an increasingly hostile environment: Breach Level Index detected a total of over 2 billion breached records during 2017.

Recurrent neural networks, which are capable of processing sequences of inputs, can be used in combination with ML techniques to create supervised learning technologies, which uncover suspicious user activity and detect up to 85% of all cyber attacks.  Startups such as Darktrace, which pairs behavioral analytics with advanced mathematics to automatically detect abnormal behavior within organizations and Cylance, which applies AI algorithms to stop malware and mitigate damage from zero-day attacks, are both working in the area of AI-powered cyber defense.

Compliance is the certification or confirmation that a person or organization meets the requirements of accepted practices, legislation, rules and regulations, standards or the terms of a contract, and there is a significant industry that upholds it.

And the volume of transaction activities flagged as potential examples of money laundering can be reduced as deep learning is used to apply increasingly sophisticated business rules to each one.  Companies working in this area include Compliance.ai, a Retch company that matches regulatory documents to a corresponding business function;

Merlon Intelligence, a global compliance technology company that supports the financial services industry to combat financial crimes, and Socure, whose patented predictive analytics platform boosts customer acceptance rates while reducing fraud and manual reviews.

While some are rightfully concerned about AI replacing people in the workplace, let’s not forget that AI technology also has the potential to vastly help employees in their work, especially those in knowledge work.

Content creation now includes any material people contribute to the online world, such as videos, ads, blog posts, white papers, infographics, and other visual or written assets.

But peer-to-peer networks are also used by cryptocurrencies, and have the potential to even solve some of the world’s most challenging problems, by collecting and analyzing large amounts of data, says Ben Hartman, CEO of Bet Capital LLC, to Entrepreneur.  Nano Vision, a startup that rewards users with cryptocurrency for their molecular data, aims to change the way we approach threats to human health, such as superbugs, infectious diseases, and cancer, among others.

Another player utilizing peer-to-peer networks and AI is Presearch, a decentralized search engine that’s powered by the community and rewards members with tokens for a more transparent search system.

It uses software to automate customer segmentation, customer data integration, and campaign management, and streamlines repetitive tasks, allowing strategic minds to get back to doing what they do best.

The software automates all the process of campaign management and optimization, making daily adjustments per ad to super-optimize campaigns and managing budgets across multiple platforms and over several different demographic and micro demographic groups per ad.

Companies Using Artificial Intelligence To Hire Employees

What if the next time you applied for a job you were interviewed by a computer?

No Such Thing As Artificial Intelligence | Two Minute Papers #60

What is happening with the neural networks in this video? You'll find the answers in these videos: 1. How Does Deep Learning Work?

AI Creates Facial Animation From Audio | Two Minute Papers #185

The paper "Audio-Driven Facial Animation by Joint End-to-End Learning of Pose and Emotion" is available here: ...

AI Learns Video Frame Interpolation | Two Minute Papers #197

The paper "Video Frame Interpolation via Adaptive Separable Convolution" and its source code is available here: ..

Artificial intelligence to produce fake news

American actor Jordan Peele is voicing Barack Obama by using AI technology and warning about how AI technology can be used to generate fake news and ...

New DeepMind AI Beats AlphaGo 100-0 | Two Minute Papers #201

The AlphaGo Zero paper "Mastering the Game of Go without Human Knowledge" is available here: ..

AI Courses by OpenCV.org

2 DAYS LEFT! Avail the 50% discount by purchasing through Kickstarter : ..

Disney's AI Learns To Render Clouds | Two Minute Papers #204

The paper "Deep Scattering: Rendering Atmospheric Clouds with Radiance-Predicting Neural Networks" is available here: ...

The Rise of the Machines – Why Automation is Different this Time

Automation in the Information Age is different. Books we used for this video: The Rise of the Robots: The Second Machine Age: ..

AI Beats Radiologists at Pneumonia Detection | Two Minute Papers #214

The paper "CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning" is available here: ...