Cloud Hosting

To be truly useful, dronesthat is, autonomous flying vehicleswill need to learn to navigate real-world weather and wind conditions.

Right now, drones are either flown under controlled conditions, with no wind, or are operated by humans using remote controls. Drones have been taught to fly in formation in the open skies, but those flights are usually conducted under ideal conditions and circumstances.

However, for drones to autonomously perform necessary but quotidian tasks, such as delivering packages or airlifting injured drivers from a traffic accident, drones must be able to adapt to wind conditions in real timerolling with the punches, meteorologically speaking.

To face this challenge, a team of engineers from Caltech has developed Neural-Fly, a deep-learning method that can help drones cope with new and unknown wind conditions in real time just by updating a few key parameters.

Neural-Fly is described in a study published on May 4 in Science Robotics. The corresponding author is Soon-Jo Chung, Bren Professor of Aerospace and Control and Dynamical Systems and Jet Propulsion Laboratory Research Scientist. Caltech graduate students Michael O'Connell (MS '18) and Guanya Shi are the co-first authors.

Neural-Fly was tested at Caltech's Center for Autonomous Systems and Technologies (CAST) using its Real Weather Wind Tunnel, a custom 10-foot-by-10-foot array of more than 1,200 tiny computer-controlled fans that allows engineers to simulate everything from a light gust to a gale.

"The issue is that the direct and specific effect of various wind conditions on aircraft dynamics, performance, and stability cannot be accurately characterized as a simple mathematical model," Chung says. "Rather than try to qualify and quantify each and every effect of turbulent and unpredictable wind conditions we often experience in air travel, we instead employ a combined approach of deep learning and adaptive control that allows the aircraft to learn from previous experiences and adapt to new conditions on the fly with stability and robustness guarantees."

Time-lapse photo shows a drone equipped with Neural-Fly maintaining a figure-eight course amid stiff winds at Caltech's Real Weather Wind Tunnel.

O'Connell adds: "We have many different models derived from fluid mechanics, but achieving the right model fidelity and tuning that model for each vehicle, wind condition, and operating mode is challenging. On the other hand, existing machine learning methods require huge amounts of data to train yet do not match state-of-the-art flight performance achieved using classical physics-based methods. Moreover, adapting an entire deep neural network in real time is a huge, if not currently impossible task."

Neural-Fly, the researchers say, gets around these challenges by using a so-called separation strategy, through which only a few parameters of the neural network must be updated in real time.

"This is achieved with our new meta-learning algorithm, which pre-trains the neural network so that only these key parameters need to be updated to effectively capture the changing environment," Shi says.

After obtaining as little as 12 minutes of flying data, autonomous quadrotor drones equipped with Neural-Fly learn how to respond to strong winds so well that their performance significantly improved (as measured by their ability to precisely follow a flight path). The error rate following that flight path is around 2.5 times to 4 times smaller compared to the current state of the art drones equipped with similar adaptive control algorithms that identify and respond to aerodynamic effects but without deep neural networks.

Out of the lab and into the sky: engineers test Neural-Fly in the open air on Caltech's campus

Neural-Fly, which was developed in collaboration with Caltech's Yisong Yue, Professor of Computing and Mathematical Sciences, and Anima Anandkumar, Bren Professor of Computing and Mathematical Sciences, is based on earlier systems known as Neural-Lander and Neural-Swarm. Neural-Lander also used a deep-learning method to track the position and speed of the drone as it landed and modify its landing trajectory and rotor speed to compensate for the rotors' backwash from the ground and achieve the smoothest possible landing; Neural-Swarm taught drones to fly autonomously in close proximity to each other.

Though landing might seem more complex than flying, Neural-Fly, unlike the earlier systems, can learn in real time. As such, it can respond to changes in wind on the fly, and it does not require tweaking after the fact. Neural-Fly performed as well in flight tests conducted outside the CAST facility as it did in the wind tunnel. Further, the team has shown that flight data gathered by an individual drone can be transferred to another drone, building a pool of knowledge for autonomous vehicles.

(L to R) Guanya Shi, Soon-Jo Chung, and Michael O'Connell, in front of the wall of fans at Caltech's Center for Autonomous Systems and Technologies

At the CAST Real Weather Wind Tunnel, test drones were tasked with flying in a pre-described figure-eight pattern while they were blasted with winds up to 12.1 meters per secondroughly 27 miles per hour, or a six on the Beaufort scale of wind speeds. This is classified as a "strong breeze" in which it would be difficult to use an umbrella. It ranks just below a "moderate gale," in which it would be difficult to move and whole trees would be swaying. This wind speed is twice as fast as the speeds encountered by the drone during neural network training, which suggests Neural-Fly could extrapolate and generalize well to unseen and harsher weather.

The drones were equipped with a standard, off-the-shelf flight control computer that is commonly used by the drone research and hobbyist community. Neural-Fly was implemented in an onboard Raspberry Pi 4 computer that is the size of a credit card and retails for around $20.

The Science Robotics paper is titled "Neural-Fly Enables Rapid Learning for Agile Flight in Strong Winds." Coauthors include Anandkumar and Yue, as well as Xichen Shi (PhD '21), and former Caltech postdoc Kamyar Azizzadenesheli, now an assistant professor of computer science at Purdue University. Funding for this research came from the Defense Advanced Research Projects Agency (DARPA) and Raytheon.

Go here to see the original:
Rapid Adaptation of Deep Learning Teaches Drones to Survive Any Weather - Caltech

First airing in 2002, The Bachelor is a titan in the world of Reality TV and has kept its most loyal viewers hooked for a full 26 seasons. To the uninitiated, the show follows 30 female contestants as they battle for the heart of a lone male bachelor, who proposes to the winner.

The contest begins the moment the women step out of a limo to meet the lead on Night One which culminates in him handing the First Impression Rose to the lady with whom he had the most initial chemistry. Over eight drama-fuelled weeks, the contestants travel to romantic destinations for their dates. At the end of each week, the lead selects one or two women for a one-on-one date, while eliminating up to five from the competition.

As self-styled mega-fans of The Bachelor, Abigail Lee and her colleagues at the University of Chicagos unofficial Department of Reality TV Engineering have picked up on several recurring characteristics in the women who tend to make it further in the competition. Overall, younger, white contestants are far more likely to succeed, with just one 30-something and one woman of color winning the leads heart in The Bachelors 20-year history a long-standing source of controversy.

The researchers are less clear on how other factors affect the contestants chances of success, such as whether they receive the First Impression Rose or are selected earlier for their first one-on-one date. Hometown and career also seem to have an unpredictable influence, though contestants with questionable job descriptions like Dog Lover, Free Spirit, and Chicken Enthusiast have rarely made it far.

For Lees team, such a diverse array of contestant parameters makes the show ripe for analysis with machine learning. In their study, Lees team compiled a dataset of contestant parameters that included all 422 contestants who participated in seasons 11 through 25. The researchers obviously encountered some adversity, as they note that they consum[ed] multiple glasses of wine per night during data collection.

Despite this setback, they used the data to train machine learning algorithms whose aim was to predict how far a given contestant will progress through the competition given her characteristics. In searching for the best algorithm, the team tried neural networks, linear regression, and random forest classification.

While the teams neural network performed the best overall in predicting the parameters of the most successful contestants, all three models were consistent with each other. This allowed them to confidently predict the characteristics of a woman with the highest probability of progressing far through the contest: 26 years of age, white, from the Northwest, works as a dancer, received her first one-on-one date on week 6, and didnt receive the First Impression Rose.

Lees team laments that The Bachelors viewership has steadily declined over the past few seasons. They blame a variety of factors, including influencer contestants (who are more concerned with growing their online following than finding true love) and the production crew increasingly meddling in the shows storylines, such as the infamous Champagne-gate of season 24.

By drawing on the insights gathered through their analysis, which the authors emphasize was done in their free time, the researchers hope that The Bachelors producers could think of new ways to shake up its format, while improving chances for contestants across a more diverse range of backgrounds, ensuring the show remains an esteemed cultural institution for years to come.

Of course, as a consolation prize, theres always Bachelor in Paradise.

View post:
Machine learning predicts who will win "The Bachelor" - Big Think

In the 60s, the Sussita was Israels first attempt at creating an Israeli car. The vehicle, made with a fiberglass body, looked like a shrunken pickup truck and was rumored to be the food of choice for camels. The vehicles carmaker, Autocars, had a very limited run and closed its doors in 1981.

Despite Israels brief moment in automotive manufacturing, the country isnt known for its car culture. And yet, since Intels acquisition of Mobileye in 2017 for $15.3 billion, Israel has been called a world leader in the autonomous vehicle marketplace, according to numerous entrepreneurs, investors, and industry experts. Why? The digitization of mobility has demanded new capabilities for vehicles and Israel has been able to provide the necessary tech sensors, LiDAR, self-driving solutions, and even, cooling tech to meet the demand. Thus, the top-tier carmakers are flocking to Israel VW, Daimler, Ford, Renault-Nissan, and many more have set up shop here.

Something similar can and should happen within the pharmaceutical industry, AION Labs CTO Yair Benita tells NoCamels.

Its been challenging to bring big pharma companies to Israel because the country doesnt necessarily have the strength in the pharma industry, he says, but now that the drug discovery process is becoming more digitized and tech is needed to leverage the influx of patient data, the government has identified an opportunity to bring to pharma to Israel, Benita says. Its very similar to what happened in the automotive industry, he adds.

Artificial intelligence (AI) and machine learning tools, in particular, are technologies that hold the key to bring about a paradigm shift in the way the pharma industry discovers and develops drugs, according to Benita and other pharma industry experts, will be speaking about the positive impact of tech on pharma as well as the challenges these companies face when dealing with pharmaceutical heavyweights on the second day of the Biomed Israel 2022 conference. The conference, which highlights the life science and technology industry, is marking its 20th anniversary this year. The three-day event, chaired by Ruti Alon, Ora Dar and Nissim Darvish, will be held from May 10-12.

Benita will chair the track titled, AI and Machine Learning Transforming Treatment, Drug Discovery, and Development. Other speakers include representatives from Israeli healthtech companiesImmunai,Quris AI,CytoReason,UKKO,Biolojic Design, as well as global experts like Eran Harary, global head, specialty R&D at Teva Pharmaceuticals, Jason Johnson,, chief data and analytics officer, SVP, at the Dana Farber Cancer Institute, and many more.

At the conference, Benita and other speakers will be discussing some of the challenges and obstacles that arise when technology enter the pharmaceutical space. Benita highlights some of these challenges below:

The first challenge is to identify the problem that is going to be solved. The problem is in the pharma space, he says. For many technology companies picking the right problem [to solve in the pharma industry] is super challenging, because often they cant intuitively understand what the big problems are for pharma without talking to them. So this is one thing were trying to solve,he explains.

The second challenge is access to data. Theres a huge amount of private data out there, Benita says, and pharma is willing to provide access to it under certain conditions. But its definitely not easy, he adds, noting that theres the challenge of validating the technology. How do you validate an algorithm? Validation occurs in the biological space or the chemical space, Benita says, And thats very challenging for any technology company.

What has been happening in recent years is that massive amounts of data are being generated across the drug discovery pipeline, thanks to AI and machine learning tech, Benita says, describing how it can be broken down into four different components.

The first one is the disease biology. We now have a huge amount of data from humans, we collect data from hospitals, and HMOs, on people. In clinical trials, we have biopsies from which we that generate a lot of data, theres a lot of data being generated on specific patients. And as you can imagine, as you look across the board, you can generate much better hypotheses that we dont generally have ever been able to do before I mean, if youre going back even 10 to 15 years, we were studying mice and animal models and not studying humans. Pharma scientists were reading papers to come up with theories, instead of looking at the data. Now were in a different position where we have all this data, and we need to come up with better hypotheses to test in the clinic. But the advantage is that the data is coming from humans, Benita explains.

Another challenge is coming up with the right business model for a technology. Many companies end up trying to develop a drug that makes sense for them but not for the drug company. And developing drugs takes too much money and too much time. So then their technology doesnt become accessible.

The last challenge for a pharma company is bringing in the right technology company rather than trying to create the technology by itself. According to Benita, pharmaceutical companies shouldnt make their own technology because its not something they know how to do. Theyre big organizations that are heavily regulated, he says. Therefore, its not a challenge for them to tackle individually, but rather to develop it together with the technology company.

The last challenge is exactly why a company like AION Labs was created, Benita tells NoCamels. The innovation lab was established from the alliance of four leading pharmaceutical companies AstraZeneca,Merck,PfizerandTeva and two leaders in the hi-tech and biotech investment sphere , respectively Amazon Web ServicesInc.(AWS)andIsrael Biotech Fund(IBF) who came together to build a lab that will spearhead the adoption of AI technologies and computational sciences while leveraging the cloud

to discover new therapies. The launch of this consortium follows the winning of a government tender to establish an Innovation Lab inDecember 2020after theIsrael Innovation Authorityidentified life sciences as a vital area for growth potential and investment.

These six partners came together to basically build companies in the pharma AI space, he says, As you can imagine, its not so trivial for for pharma companies to start working together share data-built companies. Its really difficult to start technology companies for pharma, which their main business is basically making drugs and, and testing them on humans. Its something different. These challenges are very big and it doesnt make sense for any single pharma to solve them independently. So for that reason, it made more sense to collaborate, says Benita.

Benita says AION Labs plans to create 20 to 25 new companies in the pharma AI space within the next five years, which is how long the government tender currently exists.

The pharma industry may not be like the automotive industry, but it can certainly learn a thing or two from the sector about working together to create the perfect synergy between tech and pharma.

Continued here:
How AI And Machine Learning Are Transforming Drug Discovery and Development - NoCamels - Israeli Innovation News

SAN FRANCISCO, May 03, 2022 (GLOBE NEWSWIRE) -- Today Galileo emerged from stealth with the first machine learning (ML) data intelligence platform for unstructured data that gives data scientists the ability to inspect, discover and fix critical ML data errors 10x faster across the entire ML lifecycle from pre-training to post-training to post-production. The platform is currently in private beta with the Fortune 500 and startups across multiple industries.

There are many MLOps platforms available on the market today, each fully capable of orchestrating the model lifecycle, said Bradley Shimmin, Chief Analyst of AI Platforms, Analytics and Data Management. However, when it comes to addressing the complex problem of inspecting and fixing the data -- especially for unstructured data -- many platforms still presume that enterprise practitioners work with data they already know and trust across the ML lifecycle. This couldn't be further from the truth and is one of the biggest bottlenecks for ML adoption today. What they need are tools that elevate the importance of data from the outset, putting data with a capital D back into Data Science. Galileo is tackling this critical need head on.

More than 80% of the worlds data today is unstructured (text, image, speech, etc.) and historically has been vastly untapped for ML. Recent advancements have made it easy for any data scientist to plug and play complex models for unstructured data, leading to a surge in their adoption across industries.

It is common for data scientists to use spreadsheets and Python scripts to inspect and fix their training unstructured data. Doing this data detective work consumes more than 50% of a data scientists time, is ad-hoc, manual, error prone and leads to poor data transparency across the organization, causing avoidable mispredictions and biases in production models.

Galileo takes a unique approach to this problem with just a few lines of code added by the data scientist while training a model, Galileo auto-logs the data, leverages some advanced statistical algorithms the team has created and then intelligently surfaces the models failure points with actions and integrations to immediately fix them, all within one platform. This short circuits the time taken to proactively find critical errors in ML data across training and production models from weeks today to minutes with Galileo.

Galileo goes a step further by acting as a collaborative system of record for the data scientist's training runs, bringing transparency towards how specific data and model parameter changes impact overall performance this is key for ML teams to truly be data-driven.

The motivation for Galileo came from our personal experiences at Apple, Google and Uber AI and from conversations with hundreds of ML teams working with unstructured data where we noticed that, while they have a long list of model-focused MLOps tools to choose from, the biggest bottleneck and time sink for high quality ML is always around fixing the data they work with. This is critical, but prohibitively manual, ad-hoc and slow, leading to poor model predictions and avoidable model biases creeping into production for the business, said Vikram Chatterji, co-founder and CEO of Galileo. With unstructured data across the enterprise being generated at an unprecedented scale and now rapidly leveraged for ML, we are building Galileo with the goal of being the intelligent data bench for data scientists to systematically and quickly inspect, fix and track their ML data in one place.

Galileo Founded by Engineering Leaders from Apple Google and Uber AI The co-founding team at Galileo spent more than a decade building ML products where they faced the huge challenges that ML with unstructured data present first-hand.

Galileo Focused on Data-Driven ML Research Half of the Galileo team comprises researchers from Apple, Google and Stanford AI who are focused on pushing the envelope of data-centric research that is then baked into the Galileo platform for any ML team to leverage. The other half of the team is focused on building novel systems that can perform extremely low latency in-memory computations on millions of data points using minimal system resources. This combination allows Galileo customers to get quick, intelligent data insights throughout the entire ML workflow.

Galileo Raises $5.1 Million in Seed Funding Today Galileo also announced that it has raised $5.1 million in seed funding. The Factory led the round and Anthony Goldbloom (co-founder and CEO at Kaggle) and other angel investors also participated. Company advisers include Amy Chang (Disney, P&G board member) and Pete Warden (one of the creators of TensorFlow).

Finding and fixing data errors is one of the biggest impediments for effectiveML across the enterprise. The founders of Galileo felt this pain themselves while leading ML products at Apple, Google and Uber, said Andy Jacques, investor at The Factory and Galileo board member. Galileo has built an incredible team, made product innovations across the stack and created a first of its kind ML data intelligence platform. It has been exciting to see rapid market adoption and positive reactions with one of the customers even calling the product magic!

The company plans to use the funding to hire across all departments and accelerate research and development to meet the demand of the industry for a purpose-built product to find and fix ML data blind spots across the workflow while working with unstructured data.

To read Chatterji, Sanyal and Sheths blog on ML data intelligence, simply go to: https://www.rungalileo.io/blog/introducing-ml-data-intelligence

About GalileoGalileos mission is to create data intelligence tools for unstructured data ML practitioners.With more than 80% of the world's data being unstructured and recent model advancements massively lowering the barrier to utilizing the data for enterprise ML, there is an urgent need for the right data-focused tools to build high performing models fast. Galileo is based in San Francisco and backed by The Factory. For more information, visit https://www.rungalileo.io or follow @rungalileo.

Media and Analyst Contact:Amber Rowlandamber@therowlandagency.com+1-650-814-4560

Agraphic accompanying this announcement is available at https://www.globenewswire.com/NewsRoom/AttachmentNg/a19d5c35-ff52-45e7-ba5f-ec29992ceb29

Follow this link:
Founded by Former Apple, Google and Uber AI Engineering Leaders, Galileo Launches to Give Data Scientists the Superpowers They Need for Unstructured...

Through a series of recent breakthroughs, even programmers who know close to nothing about deep learning technology can use simple, efficient tools to implement programs capable of learning from data.

If theyre lucky, they might find a book like OReillys Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, which provides concrete examples to help learners gain an intuitive understanding of the concepts and tools for building intelligent systems. (See link below for free download of chapter 2, End-to-End Machine Learning Project.)

But if someone on your tech team wants to dive even deeper into MLor any new technology, for that matterwhere can they turn?

Learning platforms give tech teams a real advantage

The future of work is becoming increasingly remote. Certainly thats been accelerated by the COVID-19 pandemic, but many studies have shown that working from home, whether full time or on a hybrid schedule, is here to stay. But when youre working remotely, who do you turn to when youre stuck? Theres no cubicle mate to rely on. Slack messages arent always instantly returned. Where can remote workers go to get the solutions they need so they can get back to work fast?

Heres where learning and development (L&D) solutions can really shineand theyre only growing brighter as we look toward how well be working tomorrow. And accommodating remote work is just the latest in a long list of transitions L&D solutions have made to help employees over the decades.

The publisher of the book noted above, OReilly, is a learning company, and for over 40 years its worked to meet tech learners where they are. But where learners are has changed over time. Thats why OReilly has transformed itselfto its book publishing enterprise, it added live global conferences, and now it has consolidated all its services into one of the most comprehensive learning platforms for tech professionals. And that transformation has resonated across the industry. Today more than 60 percent of Fortune 100 companies count on the OReilly learning platform to train their tech teams.

How do learning platforms like OReilly work?

L&D solutionsespecially those that offer certificationsare a great way for employees to learn new skills and tools that they may not have encountered during their formal education. Thats particularly true for technology professionals; often by the time they graduate, the technologies they learned in school are already becoming obsolete.

But the potential for L&D goes beyond attaining greater knowledge or advancing in title or compensation. It can also help eliminate some of the headaches all tech teams experience while tackling their daily work. For remote employees in particular, its crucial to learn in the flow of workto quickly and easily find the answers they need to overcome obstacles and get back to the task at hand.

This leads to the most recent step in OReillys evolution to meet learners where they are (both figuratively, with what theyre learning, and literally, while theyre at home). It recently added AI-enabled capabilities to create OReilly Answers. Members simply ask a question, and the platform instantly scans thousands of titles to find the best answer that solves their tech problemsometimes down to a specific line of code. Imagine all the time saved from scouring pages upon pages of books! What once took hours now takes secondsa natural language processing engine is the new cubicle mate.

The OReilly learning platform also offers thousands of live online events and training courses, where teams can ask questions and get answers from industry experts. So now if they want to attend a tech conference, they can do so from homewithout all the travel costs. Another standout feature is OReillys interactive learning scenarios and sandboxes, where teams can access safe live development environments to practice with new technologies and tools before trying to put them to work in real-world situations.

Download chapter 2 of Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow for free

Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow is just one of over 60,000 titles that are available on the OReilly learning platform. In chapter 2, End-to-End Machine Learning Project, author Aurlien Gron walks you through every step of standing up an ML project, from seeing the big picture to launching and maintaining your system. And right now OReilly is offering a free download of the chapter so potential customers can see the high-quality content available on its learning platform. Its a highly recommended read.

More:
Teaching tech teams every step of implementing a machine learning project - iTnews