My report is based on a number of articles, the first of them being entitled ‘Machine learning looks for useful data in U.S. thunderstorm reports’. It is dated September 11, 2019 and is written by Mike Krapfl of Iowa State University.
At the beginning of the article the author underlines that thunderstorms are a natural phenomenon which is not a wonder for many people. The author points out, that in our life we have all seen dozens of thunderstorms.
The National Weather Service of the USA keeps records of all storms and classifies their strength in the Storm Reports database. For a thunderstorm to be marked ‘severe’, for example, it must produce a tornado with winds greater than 58 mph.. Thunderstorms were simply classified as severe, depending, most of the time, on wind damages, such as trees down, roofs blown away or sheds pushed over, with no real measurements supporting  the estimates.
However, it’s very important  to predict severe, localized thunderstorms, and researchers need good data to help them do that.
The existing severe thunderstorm database maintained by the National Centers for Environmental Information wouldn’t be of much use for researchers looking for wind data, as wind reports were unreliable.
A BIG DATA PROBLEM
The scientists from Iowa State’s Theoretical and Applied Data Science research group decided to use computers and machine learning tools to study the reports which describe thunderstorms with severe winds.
It’s no small task as scientists will start with 12 years of severe thunderstorm reports. That’s about 180,000 of them.
Sorting through those reports raises various challenges for data researchers, because the reports are full of data collected by people, not by precise and sophisticated instruments.
Then, there are many variables, such as temperature of rising air, condensation, rainfall, lightning and more - they all have to be quantified and analyzed to understand the storms.
One of the researchers, who describes machine learning as an artificial neural network that makes connections based on the information it has available, emphasizes that computer software can handle huge amounts of storm data that would be equal to teams of people.
It’s important to note that machine learning software also does that in a very non-human way. And when people look at data they try to understand the data as human beings and therefore bring their perceptions and biases into the result of the research. One of the main reasons why the use of machine learning is so successful lies in fact that it doesn’t bring preconceived notions to its analysis of the data. Thus, it can find potential relationships that humans can’t because of their preconceptions.

Towards better forecasting.
The computers proved to make progress with the storm reports and there is also some progress in the thunderstorm wind study.
According to the scientists, their research based on the application of the machine learning could eventually lead to a new forecasting tool that predicts the possibility of severe winds, produces by a thunderstorm.

Bees’  tiny  brains  beat computers, study finds
The matter is described in the article which is more than 8 years old. It deals with an interesting research carried out by scientists at Royal Holloway, University of London, who found that bees could solve complex mathematical problems which would keep computers busy for days.
The insects can find the shortest route between flowers, using a brain the size of a grass seed. Computers solve the problem by comparing the length of all possible routes and choosing the one that is the shortest.
The scientists from Royal Holloway’s school of biological sciences studied the behavior of the bees and noted that they solve their problems every day, visiting flowers at multiple locations and, because bees use lots of energy to fly, they find a route which keeps flying to a minimum.
Using computer-controlled artificial flowers to test bee behavior, the scientists wanted to know if the insects would follow a simple route defined by the order in which they found the flowers, or look foe the shortest route.
After exploring the location of the flowers, the bees quickly learned to fly the best route for saving time and energy.
The scientists are determined to understand how the insects can solve their traveling problem without a computer.

INVASIVE SPECIES ARE A GREAT THREAT TO WILDLIFE
The problem of controlling highly invasive species without toxicants or other methods harmful to wildlife is notoriously challenging. Therefore the first experiments are aimed to discover whether biomimetric robotic fish can induce fear-related changes in mosquitofish. Researchers’ findings indicate that the application  of robotic  fish influences notably the predator’s behaviors and physiological changes associated with the loss of energy reserves, leading to lower rates of reproduction.
A team of scientists from NYU Tandon School of Engineering and the University of Western Australia have demonstrated in their research that the use of biomimetric robots can be a valuable tool in the fight against one of the world’s  most problematic invasive species, the mosquitofish.
Found in freshwater and rivers worldwide, soaring mosquitofish populations kill native fish and other wildlife which have limited options for escape. The attempts to control the species through toxicants or trapping often fail or cause harm to local wildlife.
The first study in the lab setting show that a robotic fish has a powerful, lasting impact on mosquitofish.
In some trials the robot was programmed  to incorporate with live mosquitofish and to exhibit attacks typical of predatory behavior - a rapid increase in swimming speed. The interactions between the live fish and the robot were analyzed to reveal correlations between the degree of biomimicry in the robot and the level of stress response exhibited by the live fish. Fear-related behaviors in mosquitofish include freezing (not swimming), hesitancy in exploring open spaces that are unfamiliar and other patterns.
The researchers also measured the weight and length of the fish. The decreases in weight indicate a stronger anti-predator response and result in lower energy reserves. Fish with lower reserves are less likely to survive long and devote energy toward future reproduction. This is a concrete demonstration   of the potential of robotics to solve the mosquitofish  problem.
Still, the scientists have a lot more work to establish new, effective tools to combat the spread of invasive species.
The adoption of artificial intelligence systems helps to solve a range of problems in the wildlife sciences.
So,

AI IS USED TO RECOGNIZE PRIMATE FACES IN THE WILD
Scientists at the University of Oxford have developed new artificial intelligence software to recognize and track the faces of individual chimpanzees in the wild.
Such species like chimpanzees have complex social lives and live for many years.
The researchers of the Oxford University’s Primate Models Lab., School of Anthropology, managed to employ the power of machine learning  to measure behavior of primates over the long term, for example, observing how the social interactions of a group change over several generations.
The computer model was trained using over 10 million images of wild chimpanzees in Guinea, West Africa. The new software is the first to continuously track and recognize individuals in a wide range of poses, performing with high accuracy in difficult conditions such as low lighting, poor image quality and motion blur.
The ability to closely monitor different species and populations using automated systems will be a path to finding solutions for old problems, saving hours of time and resources.

A DEEP LEARNING TECHNIQUE FOR EMOTION RECOGNITION
A team of researchers at Yonsei University and EPJL  has recently developed a new technique that can recognize emotions by analyzing people’s faces in images along with contextual features.
For several years, researchers worldwide have been trying to develop tools for automatically detecting human emotions. These tool could have numerous applications, for instance, helping doctors to identify signs of mental or neural disorders, based on atypical speech patterns, facial features, etc.
So far, the majority of techniques for recognizing emotions in images have been based on the analysis of people’s facial expressions, assuming that these expressions best convey humans’ emotional responses. However, these tools fail to achieve satisfactory performance when emotional signals in people’s faces are indistinguishable. In contrast with these approaches, human beings are able to recognize others’ emotions based not only on their facial  expressions, but also on contextual information, such as the actions they are performing, their interactions with others, etc.
The architecture developed by researchers is composed of two key subnetworks: facial features and contextual information in an image. These two types of features are combined and analyzed to predict the emotions of people in a given image.
The researchers also introduced a new dataset. Images in  this dataset portray both people’s faces and their surroundings (context). Hence, this could be more effective for training emotion recognition techniques and considerably increase the performance of the tools.

WHAT ANTS AND COMPUTERS HAVE IN COMMON
Evolutionary theorist and great myrmecologist Edward O. Wilson studied animals with complex social behavior. Leading that parade: human beings. A close second : leaf cutter ants. It might seem odd that ants, with their tiny brains, could rival humans in the sophistication of their social order.
On the surface, ants and Internet don’t seem to have much in common. But two Stanford researchers have discovered that harvester ants determine how many foragers to send out of the nest in much the same way that Internet protocols discover how much bandwidth is available for the transfer of data. The researchers are calling it the  « anternet».
Deborah Gordon, a biology professor at Stanford, has been studying ants for more than 20 years in Arizona. When she figured out how the harvester ant decided when to send out more ants to get food, she called a professor of computer science at Stanford and an expert on how files are transferred on a computer network.
At first he didn’t see any connection between his and Gordon’s work, but inspiration would soon strike. It occurred to him that it was almost the same as how Internet protocols discover how much bandwidth was available for transferring a file. The algorithm the ants were using to discover how much food there was available was similar to the one used in the Transmission Control Protocol.
It’s remarkable that ants and Internet both employ communication protocols.
It is a wonder to think that we human beings, in our most complex actions and achievements, speak a cognate language to that of the rest of nature. The discoveries like that of the «anternet» show, that we speak the same language an those countless millions of creatures speak our native tongue.
When  observing an ant colony, it is fascinating to watch hundreds of ant doing things at the same time, they function together as a little society. The researchers are trying to find the fundamental principles that underlie, for example, collective decision-making in biological systems. And what we find remarkable is, when we actually look at the algorithms used by an ant colony, when making collective decisions, the types of algorithms they use are also the types of algorithms humans use.
Ants have algorithms. If you think about an ant colony, it’s a computing device. The researchers wonder at the ways in which social insect colonies can interact, how can a colony decide between two food sources, one of which is slightly closer than the other… Do they have to measure this? Do they have to perform these computations?
The researchers also observed that ants lay a chemical trail which the other ants tend to follow. In case the source of food is closer, even if they are searching more or less at  random, ants have a higher probability of returning to the nest more quickly. Then they leave behind a stronger scent which attracts more ants.
One can also build analogies with neural systems.  Here the ants are able to move around and interact with each other, but in principle they’re  using relatively simple rules. There is a very effective collective decision-making, they function  as colonies.
The reason these ants don’t have traffic jams the way humans would is that they’re related.
When we model human crowds, we use a different algorithm entirely; individuals may minimize their own travel times, but may do so at the expense of others because they don’t necessarily care whether they cause congestion for other individuals.

COMPUTER SECURITY
In the never-ending battle to protect computer network from intruders, security experts are deploying a new defense modeled after one of nature’s hardiest creatures - the ants.
Unlike traditional security devices, which are static, these «digital ants» wander through computer networks looking for threats, such as «computer worms» - programs designed to steal information or facilitate unauthorized use of machines. When a digital ant detects a threat, it doesn’t take long for an army of ants to converge at that location, drawing the attention of human operators who step in to investigate.
The scientists who study the problems dealing with computer security and computer networks point out that in nature ants defend against threats very successfully. They can build up their defense rapidly, and then resume routine behavior quickly after an intruder has been stopped. The researchers were trying to achieve that same framework in a computer system.
Current security devices are designed to defend against all known threats at all times but intruders create software for malicious purposes. They introduce slight variations to evade computer defenses.
A research scientist at Pacific Northwest National Laboratory came up with the idea of copying ant behavior. The digital ants were tested on a network of 64 computers.
The scientists had an idea to deploy 3,000 different types of digital ants, each looking for evidence of a threat. As they move about the network, they leave digital trails modeled after the scent trails ants use in nature to guide other ants. Each time a digital ant identifies some evidence, it is programmed to leave behind a stronger scent. Stronger scent trails attract more ants, producing the swarm that marks a potential computer infection.
In their study the researchers introduced a worm into the network and the digital ants successfully found it.
The scientists claim that the new security approach is best suited for large networks, such as those found in governments, large corporations and universities.
Computer users need not worry that a swarm of digital ants will decide to take up residence in their machine by mistake. Digital ants cannot survive without  a software located at each machine, monitored by humans, who supervise the colony and maintain ultimate control.