Connect with us

Manufacturing

Researchers Create Compound Eye Based on Insects

Published

 on

Researchers Create Compound Eye Based on Insects

Researchers from Tianjin University in China have developed a newly created biologically inspired compound eye. It will be used to help scientists gain a better understanding of how insects use their compound eyes to sense objects and trajectories extremely fast. The researchers are also looking at how to use it with a camera to create 3D location systems for robots, self-driving cars, and unmanned aerial vehicles. 

The new bio-inspired compound eye was reported in the Optical Society (OSA) journal Optics Letters. It both looks like an insect as well as works like one. The compound eyes of insects consist of hundreds to thousands of ommatidia, or repeating units. Each one of them act as a separate visual receptor. 

Le Song, a member of the research team, spoke on the new project. 

“Imitating the vision system of insects has led us to believe that they might detect the trajectory of an object based on the light intensity coming from that object rather than using precise images like human vision,” Le Song said. “This motion-detection method requires less information, allowing the insect to quickly react to a threat.”

The researchers created 169 microlenses on the surface of the compound eye through a method called single point diamond turning. The microlens had a radius of around 1mm, and this created a component that was about 20mm. It was able to detect objects from a 90-degree field of view. 

One of the issues that researchers run into when creating a compound eye is that image detectors stay flat while the surface of the compound eye is curved. They got around this by placing a light guide between the curved lens and an image detector. By doing this, the team was able to enable the component to receive light from different angles uniformly. 

“This uniform light receiving ability of our bio-inspired compound eye is more similar to biological compound eyes and better imitates the biological mechanism than previous attempts at replicating a compound eye,” said Song.

When it comes to measuring 3D trajectory, the researchers put grids on each eyelet of the compound eye to help detect location. LED light sources where then placed at different distances and directions. The compound eye used an algorithm to calculate the 3D location of the LEDs using the location and intensity of the light. 

The compound eye was able to detect the 3D location of an object very rapidly. The one issue was that when the light sources were far away, the location accuracy became reduced. This could be the reason that most insects are nearsighted. 

“This design allowed us to prove that the compound eye could identify an object’s location based on its brightness instead of a complex image process,” said Song. “This highly sensitive mechanism suits the brain processing ability of insects very well and helps them avoid predators.”

The researchers believe that because this new compound eye can detect an object’s 3D location, it could be used for small robots that require fast detection from lightweight systems. This new technology can also help scientists better understand insects. 

The next step for the scientists is to put the localization algorithm into different platforms, like integrated circuits, so that the system can be used in other devices. They also want to be able to mass produce the compound eyes in order to reduce the cost. 

 

Spread the love

Deep Learning Specialization on Coursera

Investments

Microsoft Partners with Startup Graphcore to Develop AI Chips

mm

Published

on

Microsoft Partners with Startup Graphcore to Develop AI Chips

Microsoft hopes that its Azure cloud platform will catch up in popularity with Amazon and Google, so as Wired reports, it has partnered with a British startup Graphcore to come up with a new computer chip that would be able to sustain all-new artificial intelligence developments.

As Wired notes, Bristo, UK startup Graphcore “has attracted considerable attention among AI researchers—and several hundred million dollars in investment—on the promise that its chips will accelerate the computations required to make AI work.” Since its inception in 2016, this is the first time that the company is publicly coming up with its chips and testing results.

Microsoft’s invested in Graphcore in December 2018 “as a part of a $200 million funding round”, as it wants to stimulate the use of its cloud services to a growing number of customers that use AI applications.

Graphcore itself designed its chips from scratch “to support the calculations that help machines to recognize facesunderstand speechparse languagedrive cars, and train robots.” The company expects that its chips will be used by “companies running business-critical operations on AI, such as self-driving car startups, trading firms, and operations that process large quantities of video and audio, as well as those working on next-generation AI algorithms.”

According to the benchmarks published by Microsoft and Graphcore on November 13, 2019, “the chip matches or exceeds the performance of the top AI chips from Nvidia and Google using algorithms written for those rival platforms. Code is written specifically for Graphcore’s hardware maybe even more efficient.”

The two companies also stated that “certain image-processing tasks work many times faster on Graphcore’s chips,” and that “ they were able to train a popular AI model for language processing, called BERT, at rates matching those of any other existing hardware.”

Moor Insights AI chip specialist Karl Freund is of the opinion that the results of the new chip show that it is “cutting-edge but still flexible,”  and that “they’ve done a good job making it programmable,” an extremely hard thing to do.

Wired further adds that Nigel Toon, co-founder, and CEO of Graphcore, says the companies began working together a year after his company’s launch, through Microsoft Research Cambridge in the UK. He also told the publication that his company’s chips are especially well-suited to tasks that involve very large AI models or temporal data. Also, one customer in finance supposedly saw a 26-fold performance boost in an algorithm used to analyze market data thanks to Graphcore’s hardware.

Some other, smaller companies used this occasion to announce that “they are working with Graphcore chips through Azure.” This includes Citadel, which will use the chips to analyze financial data, and Qwant, a European search engine that wants the hardware to run an image-recognition algorithm known as ResNext.

Spread the love

Deep Learning Specialization on Coursera
Continue Reading

Manufacturing

Super-Compressible Material Developed Through AI

Published

on

Super-Compressible Material Developed Through AI

A new super-compressible material developed through AI by researchers at TU Delft can transform many of our everyday objects while still staying strong. The researchers did not conduct any experimental tests, and they created the material using only artificial intelligence and machine learning.

Miguel Bessa is the first author of the publication which appeared in Advanced Materials on October 14. 

“AI gives you a treasure map, and the scientist needs to find the treasure,” he said. 

Transforming Everyday Objects

Miguel Bessa, an assistant professor in materials science at TU Delft, got the inspiration to create this material after spending time at the California Institute of Technology. It was there, at the Space Structures Lab, where he observed a satellite structure that was able to open long solar sails from a small package. 

After seeing this, Bessa wanted to know if it was possible to design a super-compressible yet strong material and compress it into a small fraction of its volume. 

“If this was possible, everyday objects such as bicycles, dinner tables and umbrellas could be folded into your pocket,” he said. 

The Next Generation of Materials 

Bessa believes it is important that the next generation of materials be adaptive and multi-purpose with the capability to be altered. The way to do this is through structure-dominated materials, which are metamaterials that are able to exploit new geometries. This will allow the materials to have certain properties and functionalities that did not exist before. 

“However, metamaterial design has relied on extensive experimentation and a trial-and-error approach,” Bessa says. “We argue in favor of inverting the process by using machine learning for exploring new design possibilities, while reducing experimentation to an absolute minimum.”

“We follow a computational data-driven approach for exploring a new metamaterial concept and adapting it to different target properties, choice of base materials, length scales and manufacturing processes.

New Possibilities

Using machine learning, Bessa developed two designs that were different length scales for the super-compressible material developed through AI. They transformed brittle polymers into metamaterials which were a lot more lightweight and recoverable. The most important and impressive aspect of these new metamaterials is that they are super-compressible. The macro-scale design focuses on maximum compressibility, while the micro-scale is best for high strength and stiffness. 

Bessa argues that the most important part of the work is not the actual developed material, but it’s the new way of designing through the use of machine learning and artificial intelligence. This could open up possibilities that were unknown before. 

“The important thing is that machine learning creates an opportunity to invert the design process by shifting from experimentally guided investigations to computationally data-driven ones, even if the computer models are missing some information. The essential requisites are that ‘enough’ data about the problem of interest is available, and that the data is sufficiently accurate.”

Bessa believes in data-driven research in materials science and its ability to revolutionize and transform our way of life. 

“Data-driven science will revolutionize the way we reach new discoveries, and I can’t wait to see what the future will bring us.”

Taking Over From Start to Finish

These new developments show that there are areas that can be transformed by AI and machine learning that are not well-known. While it is proven that artificial intelligence will revolutionize machines, technologies, and almost every other aspect of society, it is not often acknowledged that it can also develop these completely on their own. There will be a point at which machine learning and AI will take over the design and development process from start to finish. It will be up to humans to instill certain mechanisms in these technologies so that they are compatible with our ways of life. 

 

Spread the love

Deep Learning Specialization on Coursera
Continue Reading

Manufacturing

Company Using AI To Automate Pizza Production

mm

Published

on

Company Using AI To Automate Pizza Production

The Seattle-based food tech company Picnic is in the process of utilizing AI to create pizzas for their customers. The deep learning algorithms used by Picnic are capable of running a pizza production line with very little oversight, analyzing the pizza at different stages with a computer-vision system.

Picnic was once called Vivid Robotics and the company has created what it dubs the first every all-purpose, automated system designed for the creation of food in the hospitality and foodservice sectors. According to TechXplore, the system is integrated with an app that customers can download and order pizzas with, customizing their toppings. The orders are given directly to the system, and the AI can oversee the creation of up to 300 12-inch or 180 18-inch pizzas every hour.

As reported by TechXplore, although the current version of the pizza production AI is just a prototype, the company expects that it will be able to create industry-ready versions of their product by the end of 2020. Picnic currently has clients lining up for their production run. The CEO of Picnic, Clayton Wood, explained that Picnic’s pizza system should cut down on labor-intensive and repetitious work, as it only requires one worker to place the dough onto the assembly line and to occasionally refill the toppings.

As many restaurants shift from dine-in operations to take-out, more food needs to be produced in a short amount of time. People are frequently frustrated and upset if their pizza takes much longer than 30 minutes. Wood thinks that Picnic can help pizza places cope with high-demand and a staffing shortage that are both creating a “great deal of stress”.

The vice president of sales at CHD Expert, a research company, Charles Chuman explained to Techxplore that there were two major factors driving the labor shortage in the food industry. One factor is the immigration constraints pushed forward by the Trump administration. The other factor is the general disruption of traditional restaurants by catering, online ordering and delivery, robotics, and AI.

Pizza itself is experiencing a partial decline in popularity in recent years as new low-carb diets have become popular and access to more diverse cuisines has expanded. While the pizza market definitely isn’t going away anytime soon, Chuman explained that the pizza industry isn’t seeing the kind of rapid growth that it was a decade ago.

Picnic has acknowledged the changing landscape of food and it is currently aiming to expand its system of automated food production out into other kinds of meals. Picnic wants to start the automated production of various pasta bowls, salads, and Mexican food using their AI-driven system.

“We think that when they’re delivering consistent, high-quality product, they’ll see increased demand for the product, as well as the savings from food waste,” explained Wood to TechXplore.

Picnic is intending to operate and upgrade their AI-driven food production system for no upfront cost, instead- charging clients a monthly fee that scales depending on the complexity of the operation. Wood expects clients to be attracted by the ease of producing constantly high-quality items.

Chuman predicts that as food production and delivery becomes quicker and easier thanks to automated systems, restaurants will need to innovate and offer different experiences or services in order to get customers to dine-in.

Picnic is far from the only company investigating how AI can be used to transform the landscape of food production and delivery. Companies like Instacart use machine-learning to predict the availability of groceries in real-time, while the start-up Sure aims to help customers decide what kind of food they should have when eating out at night. Companies like Gastrograph AI are even doing research on customer’s tastes to try and determine what kind of food and beverages they will like. Finally, companies like the Japanese start-up Calbee are using artificial intelligence to combat food waste and extend the lifespan of their products.

As AI becomes more ubiquitous and versatile, the myriad of ways it will transform the food industry will expand as well.

Spread the love

Deep Learning Specialization on Coursera
Continue Reading