Podcast: Robot That Will Save Grape Farmers Billions & How Metallic Glass Will Result In Affordable Robotics

Daniel Mitchell

fromThe Next Byte

24 Aug, 2021

Podcast: Robot That Will Save Grape Farmers Billions & How Metallic Glass Will Result In Affordable Robotics

In this episode, we talk about a collaboration between two professors at Cornell University to combat the multi-billion dollar annual loss of grapes by combating powdery mildew using AI and how a novel material developed by NASA JPL researchers can significantly reduce the cost of robotics worldwide

3D Printing

- artificial intelligence

- collaborative robots

- computer vision

- materials

In this episode, we talk about a collaboration between two professors at Cornell University to combat the multi-billion dollar annual loss of grapes by combating powdery mildew using AI and how a novel material developed by NASA JPL researchers can significantly reduce the cost of robotics worldwide. As always, you can find these and other interesting & impactful engineering articles on Wevolver.com.

EPISODE NOTES

(1:04) - Facial Recognition AI Helps Save Multibillion Dollar Grape Crop:

Powdery mildew causes multi-billion dollar losses of grape crops on a yearly basis so it is no surprise that researchers like Professor Lance Cadle-Davidson from Cornell University are working on developing more fungus resistant varieties. The process of analyzing grape leaf samples is quite cumbersome and would’ve required 6 months of manual labor had engineer and computer scientists - Professor Yu Jiang - not stepped in. Professor Jiang was able to reduce the analysis timeframe down to just 1 day by developing a robot powered by a modified facial recognition artificial intelligence model and equipped with an optical microscope.

(7:50) - Metallic Glass Gears Up For 'Cobots,' Coatings, And More:

Robots are composed of many mechanical components, the most important of which are the gears. NASA has leveraged the benefits of robotics in many missions (Mars rovers being the most famous) but the harsh atmospheric conditions have proven to be problematic for operation since the necessary lubricant for the gears has to be constantly heated up before conducting any task. NASA JPL recruited the help of Douglas Hoffman to address this problem and through his work they were able to develop bulk metallic glass gears which possess the hardness of crystals, maintain the favorable mechanical properties of metals, can be injection molded for mass manufacturing, and do not require lubrication. This breakthrough is a substantial move in the right direction concerning affordable robotics since gears account for ~50% of the price in robotic applications.

About the podcast:

Every day, some of the most innovative universities, companies, and individual technology developers share their knowledge on Wevolver. To ensure we can also provide this knowledge for the growing group of podcast listeners, we started a collaboration with two young engineers, Daniel Scott Mitchell & Farbod Moghaddam who discuss the most interesting content in this podcast series.

To learn more about this show, please visit the shows page. By following the page, you will get automatic updates by email when a new show is published.

Be sure to give us a follow and review on Apple podcasts, Spotify, and most of your favorite podcast platforms!

Use Vision AI for Accurate Livestock Monitoring at Scale

If you own, manage or operate a livestock operation, counting your animals and monitoring the movement of your livestock is a critical part of your business. See how Vision AI can help improve the process.

Sage Elliott

fromPlainsight

24 Sep, 2021

artificial intelligence

- computer vision

- machine learning

Counting sheep to fall asleep is something most of us have tried, but if you own, manage or operate a livestock operation, counting your animals and monitoring the movement of your livestock is a critical part of your business. Take a look at the above video. Do you think you can accurately count how many sheep are running down the trail in real time?

Counting sheep using computer vision seems like it should be a straightforward process, however using the computer vision tools of yesterday brought with it layers of technical complexity that required a team of data scientists to navigate. Fortunately for the livestock and agriculture industry, we here at Plainsight have developed a computer vision solution that simplifies these types of vision AI applications, while also increasing the accuracy of the count!

Most object counting systems like this require several components.

An accurate object detection model to classify and localize where the object is inside each frame of the video. This usually requires training a custom model on a labeled dataset.
An object tracking algorithm to track where each individual object moves to from frame-to-frame.
A registration zone where object detection and tracking is applied when objects enter the camera view.
A counting line that triggers the object count as each crosses it.
Tracking the movement direction across the counting line can also be important. An object that moves backward can be counted multiple times if you do not account for this possibility.
A deregistration zone where object detection and tracking can be confidently removed after the counting has happened.

In this video, the sheep enter the camera view at the bottom left in the registration zone and object detection and tracking is applied. When they walk across the line in the middle, we count them. After being counted, we

Image: By placing images on a virtual "rubber sheet," this ML approach needs little training.

One-short learning approach, inspired by human intuition, works with just one training image

Designed to dramatically reduce the amount of training data needed for an image recognition system, this one-shot approach "inspired by nativism" takes a leaf from humans' ability to intuit and abstract.

Gareth Halfacree

04 Feb, 2022

artificial intelligence

- computer vision

- machine learning

The human ability to generalize information, using abstraction to recognize everything from the shapes of letters and numbers to abstract doodles based on very limited experience, may be within reach of machine learning systems thanks to work by researchers at the Universities of Chicago and Illinois at Urbana-Champaign.

In a paper dubbed “Learning from One and One One Shot,” the research team details an approach “inspired by nativism” which allows a character recognition system to achieve human-like accuracy based on as little as one example image per class — by imagining the images, suitably processed and simplified, as being printed on a rubber sheet which can be stretched, squeezed, and bent into a range of shapes.

Abstracted learning

Machine learning systems are powerful, but suffer from one major drawback: The need for extensive training, expensive both computationally and in the datasets they ingest. A leading object recognition model may, for example, be trained on tens or even hundreds of thousands of images which have to be sourced and processed.

Approaches which look to dramatically reduce the size of training datasets are commonly referred to as “few-shot learning;” those, like this work from Haizi Yu, Igor Mineyev, Lav R. Varshney, and James Evans, which go still further to being able to generalize a single example of a given class are thus “one-shot.”

It’s a powerful idea, and one which draws inspiration from the human brain — specifically, its capacity for abstraction. “Our mind freely conceives more equivalent examples,” the researchers explain, “ignoring low-level variations in between. Babies can differentiate things quickly based on general appearance, i.e., how things ‘look in general’ (topologically, geometrically, structurally, etc.) recognizing that they may be translated, rotated, scaled, or deformed.”

Applying images to a "rubber sheet" provides a means of quantifying thei

Podcast: Robot That Will Save Grape Farmers Billions & How Metallic Glass Will Result In Affordable Robotics

In this episode, we talk about a collaboration between two professors at Cornell University to combat the multi-billion dollar annual loss of grapes by combating powdery mildew using AI and how a novel material developed by NASA JPL researchers can significantly reduce the cost of robotics worldwide

EPISODE NOTES

Abstracted learning

HyperTransformer bests rivals at few-shot learning - and takes advantage of unlabeled samples

Data Annotation for Computer Vision

Autonomous drones, triggered by a machine learning algorithm, could solve the urban feral pigeon problem

EPISODE NOTES

The Next Byte Newsletter