GRACE UNDER PRESSURE
  • Blog
  • About
  • Act Now
  • Updates
  • THESIS
  • Contact

​​

​BLOG BY GRACE C. YOUNG                                                                              
                                                                               


2018 Updates

1/22/2018

Comments

 
My Thesis: I've submitted my PhD thesis! Titled "Three Dimensional Modelling of Coral Reefs for Structural Complexity Analysis," the thesis examines the correlation between reef structural complexity and ecosystem health. 

For my thesis, I developed a new technique for creating and analyzing 3D models of underwater scenes using computer vision and machine learning. The methods are already being used by researchers in Indonesia, Madagascar, Bonaire, Cuba, Honduras, and the Maldives. I've published part of the research, and have five related publications nearing submission or under review. I hope the research will make a significant impact in our understanding of the ocean.

Ocean and Space Collaboration: I postponed submitting my thesis by two months to work in a NASA artificial intelligence accelerator last summer in California (blog post here). It was a chance to apply my knowledge of 3D modeling underwater ocean scenes to the challenge of 3D modelling near-earth asteroids. It was a fantastic opportunity that not only augmented the last chapter of my PhD thesis, but also allowed me to grow personally and professionally.

​Our team of four engineers at NASA's Frontier Development Lab used a range of machine learning techniques to automate asteroid 3D modelling. My team's results were well-received by NASA's Planetary Defense community and the tool my team developed will be implemented this year at the Arecibo Observatory to help track near-earth asteroids.

Upcoming Events: While waiting for my PhD defense in March or April, I've committed to a few speaking events, listed below.  
  • Jan 27-31 - NatGeo in London: I'll be attending NatGeo's first-ever London Explorer’s Festival. I'll share updates about current and upcoming projects with the NatGeo community, and we'll celebrate with Jane Goodall the release of her critically-acclaimed film (read more here & find a screening near you).
Picture
Washington, DC, April 10 at 7:30PM^
​General Admission
Picture
Washington, DC, April 10 10AM^ Student Matinee
Picture
Celebrating thesis submission.^
  • Feb 3 - CERN Speech in Geneva: I’ll be speaking at CERN in Geneva, Switzerland, about the “CERN for the Ocean” idea I presented in this 2015 TIME op-ed. The idea that has morphed into the a broader “CERN for the Planet.” More about my connection to CERN in previous blog post "Return to CERN." 
  • Feb 6-8 - Marine Technology Society presentations in New Orleans: I’ll be speaking in New Orleans at the Underwater Interventions conference hosted by the Marine Technology Society. With the Pisces VI submarine team, we will share updates from the build. Additionally, I'll present my thesis work on 3D modelling underwater scenes. I've attended this conference almost every year since I was 19! 
  • Feb 13 - Classroom Outreach: I’ll be hosting a live classroom session with NatGeo’s "Exploring By the Seat of Your Pants." Any classroom can sign up for free here! This session will be from our submarine workshop in Salina, Kansas, with the Pisces VI deep-sea submarine. Follow that project on Facebook and Instagram; also previous blog post "Submarines in Kansas? Yes! Pisces VI to Bring Deep Sea Research to the Masses."
  • Feb 26-27 - MIT Media Lab: I'm thrilled that Katy Croff Bell is launching the Open Ocean Initiative at MIT Media Lab. It will "design and deploy new ways to understand the ocean and connect people to it, empowering a global community of explorers." I'll be at one of the first events "Here Be Dragons" in Boston this February. Most of the talks will be live streamed from MIT or the New England Aquarium. More information here.
  • April 10 - Join me for "Extreme Oceans" Presentations in DC (a Student Matinee at 10AM,  and a General Admission at 7:30PM): I’ll be speaking at NatGeo Headquarters in Washington D.C. about why I care so much about the ocean, how my interests developed, and more broadly about the future of our ocean, the technology and policies we need, and how everyone can get involved. I'll give both a students-only matinee and an evening talk open to the public. Tickets are available here through NatGeo.  Friend Anand Varma (who did the awesome slo-mo story of on hummingbirds you might've seen!) will be speaking a few days later on April 13. 

Comments

Summer at NASA - Update

8/4/2017

Comments

 
So far so good! My previous blog post explains why I'm at NASA this summer. In short, I'm still 'Team Ocean' (of course!), but the 3D shape modelling techniques developed for my PhD on coral reefs have direct application for NASA's research on near-Earth asteroids (and vise versa). It's been a fantastic collaboration. Here are more details about what we're doing and why. 

What We'RE Doing and Why 

Picture
<- Explaining our work during a Facebook live event for SETI (here on Facebook; it's been viewed by >30k!). 
NASA's Frontier Development Lab (FDL) is an experimental tool in NASA’s innovation portfolio that emphasizes artificial intelligence, inter-disciplinary approaches, rapid iteration, and teamwork to produce significant breakthroughs useful to the space program.

​This summer, four of us at NASA FDL are creating 3D models of asteroids. Our core team comprises two planetary scientists (Agata Rozek and Sean Marshall), two machine learning engineers (Adam Cobb and me), plus mentors from both disciplines (Chedy Raissi,  Michael Busch,  and Yarin Gal). We’re creating the 3D models from radar data. It's a difficult computational problem, but knowing an asteroid’s 3D shape helps us predict its future trajectory (/whether it will collide with Earth!). 

The formal introduction to our problem reads as follows: 
​Delay-Doppler radar imaging is a powerful technique to characterize the trajectories, shapes, and spin states of near-Earth asteroids and has yielded detailed models of dozens of objects. Since the 1990s, delay-Doppler images have been analyzed using the SHAPE software developed originally by R. S. Hudson and S. J. Ostro [1, 2]. SHAPE normally performs sequential single-parameter fitting. Recently, multiple-parameter fitting algorithms have been shown to more efficiently invert delay-Doppler data sets, thus decreasing runtime while improving accuracy [3]. However, reconstructing asteroid shapes and spins from radar data is still, like many inverse problems, a computationally intensive task that requires extensive human oversight. The FDL 2016 team explored two new techniques to better automate delay-Doppler shape modeling: Bayesian optimization [4] and deep generative models [5]. The FDL 2017 team is refining that work and exploring new directions for more quickly and accurately generating 3D models of near-Earth asteroids from delay-Doppler images.
It took me a bit to understand exactly what our goals and motivations were. The most common questions my friends ask are, “What are you doing?” and “Why?” My short answer: We're generating 3D models of asteroids from radar data so that we can better determine asteroids' physical properties and orbital trajectories. There are over 16,000 known near-Earth objects, and on average 35 new ones each week. It's too much data to keep up with without sophisticated data analysis techniques, so we're  using machine learning to speed up and automate the process of generating 3D models from radar data of asteroids.

Still #TeamOcean

I'm also interested in the task of 3D modelling asteroids because the techniques can be applied to 3D modelling coral reefs, the topic of my thesis, as further discussed in my first post about NASA.

Preliminary Results

1

Picture
Result 1A: Delay-Dopler images (example above) are converted into 3D models of asteroids (example at right).
Picture
Result 1B: Last year a team trained a neural network to generate 3D asteroid shapes in the form of voxels (cube-like 3D pixels). We've developed triangular meshes from those voxels, and have smoothed the 3D shapes so that they better resemble asteroids. We'll be feeding a set of synthetic radar shapes into a deep neural network to train the network. For more details, stay tuned for our presentation on August 17th in Silicon Valley. ​​​

2

Picture
Result 2: We wrote a script that that finds signals in sets of delay-Doppler radar images. This quickens pre-processing of the data. The script intelligently masks the signal from the noise in an image using a density-based clustering (DBSCAN) algorithm.

3

Picture
Result 3: We also wrote a script that estimates the spin state of an asteroid from available data. That data can be radar data, optical or light curve data, or any of the input sources used by existing 3D modeling software for asteroids called SHAPE. It quickly and efficiently estimates spin states by performing Bayesian optimization on a spherical coordinate system. Already processing time has gone down from 3 days to 4 hours (and getting faster!).
More details will be in our final presentation and report at the end of the summer. Register here if you'd like to attend our final presentation in Santa Clara, California. 

References: 
​[1] R. Scott Hudson. Three-dimensional reconstruction of asteroids from radar observations. Remote Sensing Reviews 8, 195–203, 1993.
[2] Christopher Magri, Michael C. Nolan, Steven J. Ostro, and Jon D. Giorgini. A radar survey of main-belt asteroids: Arecibo observations of 55 objects during 1999-2003. Icarus 186, 126–151, 2007. 
[3] Adam H. Greenberg and Jean-Luc Margot. Improved algorithms for radar-based reconstruction of asteroid shapes. The Astronomical Journal 150(4), 114, 2015. 
[4] Jonas Mockus. Bayesian Heuristic Approach to Discrete and Global Optimization: Algorithms, Visualization, Software, and Applications. Springer-Verlag, Berlin, Heidelberg, 2010. 
[5] Ruslan Salakhutdinov. Learning deep generative models. Annual Review of Statistics and Its Applications 2, 361–385, 2015. 
​[6] Shane Carr, Roman Garnett, and Cynthia Lo. BASC: applying Bayesian optimization to the search for global minima on potential energy surfaces. International Conference on Machine Learning. 2016.

This post is modified from the original published on the NASA FDL page (here). All work was developed while at NASA Frontier Development Lab, working with Agata Rozek, Sean Marshall, Adam Cobb, Justin Havlovitz, Chedy Raissi,  Michael Busch,  and Yarin Gal.  

UPDATE - 12 Sept 17

My colleague Adam just posted his perspective on the project. Read his blog post here. 

Update - 20 Nov 17

The video of our final presentation at Intel Headquarters is live! It's on YouTube at this link.
Picture

Update - Jan 2018

The results from our team of four engineers and scientists were well-received by NASA's Planetary Defense Community. The tool we developed will be implemented this year at the Arecibo Observatory to help track near-earth asteroids.

Related blog posts:
  • Looking up! NASA this Summer
  • Aquarius Day 3: Met an Astronaut Underwater
Comments

LOOKING UPWARDS THIS SUMMER -- INVITE TO NASA

5/25/2017

Comments

 
Excited to announce that I’ve been offered a research position this summer at NASA’s Frontier Development Lab in Mountain View, CA. I’ll be working alongside other scientists and engineers for eight weeks in an intensive research accelerator focused on artificial intelligence.
Hosted by the SETI Institute and NASA Ames and supported by leaders in AI from the private sector, such as IBM, Nvidia, Autodesk and Miso Technologies, FDL brings together teams of experts in the physical sciences and specialists in data science and machine learning for an intense 8-week concentrated study on topics important to NASA – and to humanity’s future. The format encourages rapid iteration and prototyping to create outputs with meaningful application, papers and conference posters.
They've asked me to develop a project with cohorts on near-earth object 3D shape modelling and lunar water detection, topics that directly relate to my thesis (minus the "lunar" part and replace "near-earth" with "underwater" of course!). It's a fantastic opportunity to develop skills and learn from NASA in ways that will not only further develop our underwater 3D modelling techniques, but also vise versa; they'll learn from our research techniques. 
Picture
Picture
Image: G. Young Credit Mission 31.

Still Team Ocean

Ocean and space are commonly pitted against each other – upward vs downward – astronaut vs aquanaut, etc. ​
Fabien Cousteau and Liz Magee from Mission 31 recently debated in a Science Throwdown: Space vs. Sea at the American Museum of Natural History, and Nat Geo 2011 Explorer of the Year Kenny Broad plans to highlight the topic in his Red Planet vs Blue Planet session at the National Geographic Explorer’s Festival this June.

​I delve into the debate’s history in my undergraduate thesis 
Missiles and Misconceptions: Why We Know More About the Dark Side of the Moon Than the Depths of the Ocean, TEDx Talk Why We Know More About the Moon Than the Depths of the Ocean, and in an interview for an upcoming PBS documentary based on Ben Hellwarth’s awesome book SEALAB.
Picture
It’s not an either-or debate, however. We can explore both.
​

I’m looking forward to this summer opportunity at NASA for a number of reasons. I’m eager to learn from NASA methods that will enhance our underwater 3D modeling techniques while sharing what we’ve learned underwater. The experience will also provide another perspective on how public-private partnerships can work effectively to achieve defined research objectives. I believe public-private partnerships like NASA FDL and what I observed at CERN OpenLab, are the key to tackling our most urgent  ocean research objectives---a vision I outlined in my 2015 TIME op-ed. Finally, I’m excited to spend weekends diving, surfing, and reuniting with West Coast friends.
Picture
Image: G. Young credit Mission 31

Many astronauts are also aquanauts, but most people don’t realize more people have been to space than have lived underwater! Several acquaintances work in both arenas: Pisces VI submarine owner Scott Waters is also on a space advisory board. FDL founding member, Jordan McRae, invented OctoTalk, a system for divers to transmit voice communications underwater. Jonathan Knowles, who is involved with FDL through Autodesk, also advises The Hydrous, an organization that 3D models corals around the world.
​
That said, I’m committed to Team Ocean. I’m personally drawn to water, a connection many humans have (see Wallace Nichols' book 
Blue Mind: The Surprising Science That Shows How Being Near, In, On, or Under Water Can Make You Happier, Healthier, More Connected, and Better at What You Do.).

Moreover, engineering-wise, while space may seem more exotic, I’d argue the ocean, particularly the deep ocean, is a more challenging work environment considering that E&M waves, upon which wifi, GPS, and many modern innovations are based, do not work, and salt water kills electronics. We’re forced to innovate.

(Thanks Somerville (Oxford) for the article with video!)
Comments

AQUARIUS DAY 3: MET AN ASTRONAUT, UNDERSEA!

6/18/2014

 
Picture

​Aquaman actor Adrian Grenier waved hi to me  through the viewport yesterday. 

Day 3 already! Oh man. Time is flying by. Even Fabien thinks so. As of today, he, Otter and Ryan have been here 19 days, making it one day longer than the longest previous stay in Aquarius. Just incredible. 

6:00 AM

When I rolled out of bed, every one was already up and at'um, eating breakfast, chatting, and prepping gear for the day. I started pulling on my wetsuit at about 6:30am, meanwhile reviewing with Liz our plan for the dive. 

7:00 AM

Liz and I started our dive. Before every dive we spend about 10 minutes going over a checklist with Otter, who reads the checklist, and Ryan, who helps us put the Kirby Morgan dive helmet on. (My dive helmet post with Lady Gaga is here.) In going over the checklist, we check the functionality of various valves and emergency back-ups on the helmet. Every time everything works, as far as I know. The checklist can seem tedious, but I realize how important it is. I started thinking back to Atul Gawande's book The Checklist Manifesto:  How to Get Things Right. I remember hearing that the medical profession, which suffers from human error, should really use checklists, but doctors just don't. I see why. They're annoying. And everything seems to work. But then again, the checklist seems like the only way to ensure all is as it's suppose to be. Not only do we catch anything not functioning, but routinely checking all the valves, etc. puts our mind at ease. I never have that feeling of, "Am I forgetting something...." I just focus on my dive, and I feel like the master of my equipment. 

This morning of our dive we collected samples in jars from the plankton traps (described in this previous post), and then stored the jars in a mesh bag that Northeastern divers will pick up in the afternoon. We also adjusted a tripod around a giant barrel sponge. The tripod suspends a sensor above the sponge, and the sensors measures the sponge metabolism. 

9:00 AM

We returned to Aquarius from our first dive. I ate a late breakfast and change both the the lens and set-up on the Edgertronic camera in case we have extra time to use it on our afternoon dive. 
Picture
Good morning from The Weather Channel.
Picture
Exploring under the sea.

Met an astronaut, in the water!

On our second dive of the day, from noon to 3pm, we greeted astronaut Clay Anderson in the water. It was unreal. Fabien, Liz and I all shook his hand as he stepped onto the "porch" of Aquarius. We then gave him a tour around the house. A video of his visit is here. There's an interesting New York Times' article about the space and sea: Cosmic Connections in the Deep Sea. Living undersea, I have a new appreciation for our planet as a whole. Both astronauts and aquanauts are willing to live in an alien environment to explore the unknown, which links our cosmic work, whether it be in space or ocean. We are still looking for life in outer space, but there's so much undersea. 

During our afternoon dive, Liz and I also used a plankton tow to collect more plankton samples. You can read more about the research we are doing with the Northeastern topside team at their blog, specifically Amanda's post about zooplankton. 
Picture
Exploring the unknown cosmos links space and ocean scientists. Credit NYT, Elwood H. Smith

Astronauts and Aquanauts on Jacques B'Day

Space and ocean and space explorers alike remembered Jacques Cousteau, the great explorer, on his birthday this month. See my post Happy Birthday  Jacques Cousteau! for details of the celebration.
Picture
Mission 31 Aquanauts on Jacques Cousteau's birthday 2014. It seems aquanuts & astronauts share a common desire to explore the unknown.
Picture
NASA_ NEEMO9 Astronauts also celebrating on Jacques Cousteau's birthday 2014.

OUTREACH, Great Questions

Picture
Skyping with classrooms ... the students have great questions. Credit Elizabeth ten Grotenhuis. 

In between research dives, we are Skyping with different groups around the world, spreading the word about how exciting and important the oceans are. Yesterday, I Skyped with the AAT Project and the Birches School (read about my fall visit there) and from the research boat last week. Birches' kindergarten through third grade students asked:
  • What did you eat for breakfast? 
  • What fish have you seen? 
  • What surprised you? 
  • Do you miss your family? (Of course!) 
  • Is it cool?? (Yes!!) 
  • Grace, our chickens hatched and we are raising tadpoles! (Awww...)
There are more great adorable questions from this group in the fall.

It's BIZARRE... 

To be able to see myself working, on the Mission 31 Live Cam. 
Picture
E.g., this is me the first night here, staying up late. At MIT Commencement, President Reif said, “MIT gradutes are naturally nocturnal, especially when chasing a fresh idea.” I guess it's true.  

More Grace Under Pressure blog at Aquarius Day 4: Science and Ballet Art Undersea. 
    Picture

    Author

    Grace Young  (B.S., MIT, Ph.D, Oxford) is an ocean engineer, aquanaut, and explorer currently working at X. She lived underwater as a scientist and engineer on Fabian Cousteau’s Mission 31, and is a National Geographic Explorer. 

    Blog Highlights: 
    1. No Engineer is an Island
    2. Mission 31 Highlights
    3. Sailing Across the Atlantic 
    ​3. Return to CERN

    Tweets by @grace_h2o
    ​INSTAGRAM

    RSS Feed

    Categories

    All
    Arts & Science
    Conservation
    Coral Research Mission
    Edgertronic
    Marine Robotics
    Mission 31
    Mission 31 Training
    Ocean Reports & Facts
    Ocean & Space Science
    Outreach
    Research
    Sailing & Adventures
    Sea Creatures

Powered by Create your own unique website with customizable templates.