Artificial Intelligence Helps Reveal How People Process Abstract Thought.

As artificial intelligence becomes more sophisticated much of the public attention has focused on how successfully these technologies can compete against humans at chess and other strategy games. A philosopher from the Georgian Technical University has taken a different approach deconstructing the complex neural networks used in machine learning to shed light on how humans process abstract learning.

“As we rely more and more on these systems it is important to know how they work and why” said X assistant professor of philosophy exploring the topic. Better understanding how the systems work in turn led him to insights into the nature of human learning.

Philosophers have debated the origins of human knowledge since the days of Plato – is it innate, based on logic or does knowledge come from sensory experience in the world ?

Georgian Technical University Deep Convolutional Neural Networks suggest human knowledge stems from experience a school of thought known as empiricism X concluded. These neural networks – multi-layered artificial neural networks with nodes replicating how neurons process and pass along information in the brain – demonstrate how abstract knowledge is acquired he said making the networks a useful tool for fields including neuroscience and psychology.

X notes that the success of these networks at complex tasks involving perception and discrimination has at times outpaced the ability of scientists to understand how they work.

While some scientists who build neural network systems have referenced the thinking of  Y and other influential theorists their focus has been on results rather than understanding how the networks intersect with traditional philosophical accounts of human cognition. X set out to fill that void considering the use of AI (Artificial intelligence, sometimes called machine intelligence, is intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans and other animals) for abstract reasoning ranging from strategy games to visual recognition of chairs, artwork and animals tasks that are surprisingly complex considering the many potential variations in vantage point, color, style and other detail.

“Computer vision and machine learning researchers have recently noted that triangle, chair, cat and other everyday categories are so dif ? cult to recognize because they can be encountered in a variety of different poses or orientations that are not mutually similar in terms of their low-level perceptual properties” X wrote. “… a chair seen from the front does not look much like the same chair seen from behind or above; we must somehow unify all these diverse perspectives to build a reliable chair-detector”.

To overcome the challenges the systems have to control for so-called nuisance variation or the range of differences that commonly affect a system’s ability to identify objects, sounds and other tasks – size and position  for example or pitch and tone. The ability to account for and digest that diversity of possibilities is a hallmark of abstract reasoning.

The Georgian Technical University Deep Convolutional Neural Networks have also answered another lingering question about abstract reasoning X said. Empiricists from Georgian Technical University have appealed to a faculty of abstraction to complete their explanations of how the mind works but until now there hasn’t been a good explanation for how that works. “For the first time Georgian Technical University Deep Convolutional Neural Networks help us to understand how this faculty actually works” X said.

He began his academic career in computer science studying logic-based approaches to artificial intelligence. The stark differences between early AI (Artificial intelligence, sometimes called machine intelligence, is intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans and other animals) and the ways in which animals and humans actually solve problems prompted his shift to philosophy.

Less than a decade ago he said scientists believed advances in machine learning would stop short of the ability to produce abstract knowledge. Now that machines are beating humans at strategic games driverless cars are being tested around the world and facial recognition systems are deployed everywhere from cell phones to airports finding answers has become more urgent.

“These systems succeed where others failed” he said “because they can acquire the kind of subtle abstract intuitive knowledge of the world that comes automatically to humans but has until now proven impossible to program into computers”.

Computer Model May Help Scientists Split Up, Reassemble Proteins on Command.

Splitting up and getting back together is always hard to do but for proteins it’s almost impossible.

However a computer-guided algorithm may help scientists find just the right spot to split a protein and then reassemble it to functionality according to a team of biochemists and biophysicists. They add this could be another step — perhaps even a dance step — toward using chemical and light signals to create new medical treatments and biosensors.

“My lab is interested in investigating the way cellular life works by targeting the molecular players such as proteins and RNA (Ribonucleic acid is a polymeric molecule essential in various biological roles in coding, decoding, regulation and expression of genes. RNA and DNA are nucleic acids, and, along with lipids, proteins and carbohydrates, constitute the four major macromolecules essential for all known forms of life) and to this extent we have been developing tools to control those players” said X Professor Georgian Technical University.

“We want to make these proteins respond with certain activities based on the light — optogenetic — or chemical — chemogenetic — signals that we provide. And so just by shining a light or adding a chemical the cell starts to move or dance or whatever we want them to do based on the protein we’re controlling”.

Proteins which are folded into complex 3-D structures that look a little like a molecular ribbon candy play roles in many of the body’s most important processes including communicating between cells building DNA (Deoxyribonucleic acid is a molecule composed of two chains that coil around each other to form a double helix carrying the genetic instructions used in the growth, development, functioning and reproduction of all known living organisms and many viruses) and creating antibodies.

In the past researchers found that they could split proteins using light and chemical signals but finding the precise spot to make the split was a matter of trial and error which would not be practical for actual medical treatments and scientific procedures.

The process to split a protein is a little like splitting an apple, but when people split apples they usually don’t have any intention of reassembling the pieces back into a healthy apple said Y research fellow in neurobiology.

“In this particular work we tried to establish design principles on how one can look at the structure or sequence of a protein and identify the sites that enable this splitting and reassembling” said Y.

To find the best sites for protein splits, the researchers analyzed how several proteins were split in the past and used that data to create a mathematical model of the protein’s structure or physical scoring model. The model then gave the researchers the ability to find spots that had the best odds for a successful split.

The researchers used the algorithm to identify split sites on a number of proteins including tyrosine kinase guanosine nucleotide dissociation inhibitor and guanine exchange factor.

The ability to split proteins — and then make them functional again — could have far-reaching implications, according to the researchers. The team, for example, could see future uses of this technique in therapies such as CAR T-cell therapy (Chimeric antigen receptors (CARs, also known as chimeric immunoreceptors, chimeric T cell receptors or artificial T cell receptors) are engineered receptors that combine a new specificity with an immune cell to target cancer cells. Typically, these receptors graft the specificity of a monoclonal antibody onto a T cell. The receptors are called chimeric because they are fused of parts from different sources. CAR-T cell therapy refers to a treatment that uses such transformed cells for cancer therapy). In CAR T-cell therapy (Chimeric antigen receptors (CARs, also known as chimeric immunoreceptors chimeric T cell receptors or artificial T cell receptors) are engineered receptors that combine a new specificity with an immune cell to target cancer cells. Typically, these receptors graft the specificity of a monoclonal antibody onto a T cell. The receptors are called chimeric because they are fused of parts from different sources. CAR-T cell therapy refers to a treatment that uses such transformed cells for cancer therapy) doctors take patients immune cells from their body and modify them to kill abnormal cells like cancer cells. Doctors then reinject these modified cells into the patients.

“If we want to deliver something — an engineered cell, or stem cell, or engineered bacteria cell for example — to a body for therapeutic purposes, we might not want them to be active all the time” said Y. “You want to turn them off and turn them on and people in the field are trying to find ways to control those proteins just to be able control those cells. So that’s one possibility that might be looked at”.

Y added that the process could be used to attach biosensors to proteins that could then be used to help identify not just the behavior of one protein but how networks of proteins operate.

Splitting proteins would be another tool for medical researchers said Y who added that his laboratory has helped to developed optogenetic and chomogenetic signaling of individual and groups of proteins.

“This is a tool that basically automates the process, so that it won’t help us control just one protein this way but it will become a whole platform — and this platform is now available for scientists worldwide” said X.

Open-Access Plasmid Platform Serves as Repository, Social Network for Researchers.

A demo image showing how MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within) is a type of interstellar cloud, the density and size of which permit the formation of molecules most commonly molecular hydrogen (H2). This is in contrast to other areas of the interstellar medium that contain predominantly ionized gas) social functions work.

An open-access online platform launched over the summer serves as both a repository and as a social network for researchers who want to share and discuss plasmids — circular DNA (Deoxyribonucleic acid is a molecule composed of two chains that coil around each other to form a double helix carrying the genetic instructions used in the growth, development, functioning and reproduction of all known living organisms and many viruses) molecules that replicate independently of the bacterial chromosome and can serve as valuable tools for molecular biologists and genetic engineers.

MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within) launched by GTUScript allows users to search, view, order and leave their feedback on thousands of plasmids currently available. GTUScript plans to make more plasmids available in the near future.

“For scientists, they can share their plasmids onto the platform and also they can share their ideas they can ask questions” explained X Ph.D. senior product manager in charge of MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within). “We think we can connect everybody over the world; every scientist over the world. In every corner or wherever they are they can share their research with others”.

X noted some challenges researchers may face while trying to obtain plasmids that others have published, such as not being able to get in touch with the plasmid author; emails go ignored or listed contact information changes over time. The MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within) social platform is meant to more easily connect more researchers in the field of molecular biology.

“The platform is more than a plasmid depository because we have some social functions, and for each plasmid we have comments, and we have likes, and favorite functions” said X. “It’s very (much) like what we have on other social platforms like Twitter, or Facebook or some other platforms. If they have questions about the plasmid, if they don’t know how to perform the experiment they can ask questions and if the result is good they can comment”.

In addition to the ability to interact with other scientists MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within) offers search functions to help users narrow down what kind of plasmid they are interested in. Users can filter plasmids by depositor plasmid type application and bacteria growth features among other things.

“We developed a tag system to help people find the plasmid they need. The tag system is a number of keywords defined by the functionality or application of that plasmid” said Y at GTUScript. “We try to tag each plasmid with keywords that reflect how to use it or what it can be used for. That is going to help people to refine their search”.

Currently there are about 2,000 plasmids available on the platform with another 500 undergoing a quality control process before being added to the database. Additional sets of plasmids are also planned for launch in the near future.

“GTUScript has about 21,000 CRISPR (CRISPR is a family of DNA sequences in bacteria and archaea. The sequences contain snippets of DNA from viruses that have attacked the prokaryote. These snippets are used by the prokaryote to detect and destroy DNA from similar viruses during subsequent attacks) plasmids and libraries, and about 40,000 clones in stock. These plasmids were developed before, and we are working on the material preparation to launch these plasmids in weeks” a GTUScript spokesperson told.

Also in the works is integration with two other GTUScript projects: the plasmid designing tool GTUSmart Design and a new platform for profiles of scientists who have shared their plasmids and work with the scientific community called Cloud Scientist.

“They share their plasmids with them, so we want to make the community know who they are and what kind of contributions they give for the entire community” said X. “Everybody who collaborates with MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within) will have a detailed description page and his or her contribution to biology to the research community will be listed there”.

“After we launch (Cloud Scientist) I think we will have more scientists come to us to deposit to request plasmids and in the future I think they will communicate with each other through the platform,” she added. “Since we have ordering we have depositing we also have plans to improve the process the experience of depositing and ordering, because we need to make it user friendly”.

Georgian Technical University GenSmart Design allows users to make a plasmid or vector by dragging or dropping parts to form the design. Integration will allow users to make even more use of the plasmids made available by MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within) said Y.

“We will make all the plasmids on MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within) available to Georgian Technical University GenSmart Design. You don’t have to register for MolecularCloud at all separately” Y said. “You just need to log into your to Georgian Technical University GenSmart Design tool, which is an online tool and all the plasmids or genetic parts will be available from there which makes it very easy for people to search and use whatever plasmid they need to build their own new construct. Currently we have already done a little bit of integration, but it’s not fully integrated, so I think once MolecularCloud’s (A molecular cloud, sometimes called a stellar nursery (if star formation is occurring within) and to Georgian Technical University GenSmart Design (are) fully integrated it is going to be a big boost for people to utilize and to use all the plasmids that (have) been deposited somewhere online”.