This month I'm making time to learn new skills. Continuing with my interest in machine learning and data science has to be somewhat extracurricular, because my current research project doesn’t really call for it. Well, I’m doing “regression analysis” by fitting models to data, but I’m not really practicing any of the ML techniques that I learned about last August at the data science summer school I attended, e.g., classification (incidentally, S2DS is currently accepting applications for this year’s class).
To learn some new skills, I’ve been spending 1 hour per day with the book Statistics, Data Mining, and Machine Learning in Astronomy by Z. Ivezic, A. Connolly, J. VanderPlas, and A. Gray (note they use the Oxford comma - let that be a lesson to you). This book is also known as AstroML.
The book is divided into two parts, and I’ve decided to work through it by bouncing between them. I’m not sure it is the most efficient way to work through it, but I think it will keep my interest longer. I’ve worked through Chapters 1 & 2, and now I’m working through Chapter 6. After that, I'll got back to Chapter 3, and so on. A research partner of mine and I have some ideas that we want to use ML techniques for. Hopefully this book will teach us how to do that.
There is a website that accompanies the book, astroml.org, and it contains the code for the figures found within. We are working through those, trying to understand as best we can, so we can implement similar techniques in our research.
This project has two main goals for me. The first, as stated above, is that it is interesting and will lead to some interesting research. The second is that it will really help develop my professional skills. I’m still waiting to hear back from a couple positions that I applied to, but I’m also starting to diversify my applications. In February, I’ve applied to 11 positions in industry so far - that’s in addition to the 20+ academic positions that I applied to during the Nov-Jan period. The academic application period is slowing down now, so if I really want to come back to the U.S. anytime soon, I’ll need to diversify.
That’s what’s new with me,
Kyle
Friday, February 20, 2015
Sunday, January 25, 2015
A Collaborative Visit
I’ve just returned from a visit to Middlebury College in Vermont, where I gave a talk and visited with a collaborator.
During my graduate work, I ended up with some extra data. I had applied for observations of a sample of red quasars several years in a row and was denied time. On the third time, I/we were granted time to observe these objects with NASA’s Infrared Telescope Facility on Mauna Kea in Hawaii. This was really great, but by that time I was ready to graduate. I went observing and took the data, but they did not make it into my Ph.D. thesis. Instead, my advisor and I decided to give the data to someone who is an expert in red quasars, and she was glad to have the data.
Since then the data have been reduced and our collaborator let a student work on some of the analysis. The student calculated black hole masses from the spectra that I observed, plus a series of other spectra that were not mine in origin. And since I am currently visiting Yale University in New Haven, CT, I thought it would be easy to drive up to Middlebury, VT for a visit and status update on the infrared spectra. It turns out the student did a lot of great work, and we are looking forward to turning her work into a publication.
While at Middlebury College I was asked to give a talk. Academic astronomy talks are usually somewhere between 30 min and 1 hour, depending on the level of formality. The one exception is that at the American Astronomical Society meetings, you only get 5 minutes to present your current research. Those meetings are just so big that there is no way to give people more time. Anyway, I put together my slides during the couple days before actually going to Middlebury. I was a bit hesitant at first, because I haven’t given a presentation recently and I felt out of practice. But, that also was a strong motivator.
It turned out really well. I got a lot of positive feedback from several people in the audience, and I got a lot of questions after the talk. I was very happy to receive the feedback, because I was a little nervous going into the presentation.
After the talk, we had a little group meeting between the collaborators on the red quasar project. It went really well, and I was impressed by the amount of work that the student had done, who is an undergraduate. And those discussions naturally led into discussions about a future project. I will write a proposal to observe these objects with a facility in Chile, which I think will be really useful to learn about the physics that is going on in this sample and will tell us about black hole - galaxy evolution.
So that’s my update for the week. It was a good week.
Oh yeah - Airbnb and poutine are really good - although, not necessarily related.
During my graduate work, I ended up with some extra data. I had applied for observations of a sample of red quasars several years in a row and was denied time. On the third time, I/we were granted time to observe these objects with NASA’s Infrared Telescope Facility on Mauna Kea in Hawaii. This was really great, but by that time I was ready to graduate. I went observing and took the data, but they did not make it into my Ph.D. thesis. Instead, my advisor and I decided to give the data to someone who is an expert in red quasars, and she was glad to have the data.
Since then the data have been reduced and our collaborator let a student work on some of the analysis. The student calculated black hole masses from the spectra that I observed, plus a series of other spectra that were not mine in origin. And since I am currently visiting Yale University in New Haven, CT, I thought it would be easy to drive up to Middlebury, VT for a visit and status update on the infrared spectra. It turns out the student did a lot of great work, and we are looking forward to turning her work into a publication.
While at Middlebury College I was asked to give a talk. Academic astronomy talks are usually somewhere between 30 min and 1 hour, depending on the level of formality. The one exception is that at the American Astronomical Society meetings, you only get 5 minutes to present your current research. Those meetings are just so big that there is no way to give people more time. Anyway, I put together my slides during the couple days before actually going to Middlebury. I was a bit hesitant at first, because I haven’t given a presentation recently and I felt out of practice. But, that also was a strong motivator.
It turned out really well. I got a lot of positive feedback from several people in the audience, and I got a lot of questions after the talk. I was very happy to receive the feedback, because I was a little nervous going into the presentation.
After the talk, we had a little group meeting between the collaborators on the red quasar project. It went really well, and I was impressed by the amount of work that the student had done, who is an undergraduate. And those discussions naturally led into discussions about a future project. I will write a proposal to observe these objects with a facility in Chile, which I think will be really useful to learn about the physics that is going on in this sample and will tell us about black hole - galaxy evolution.
So that’s my update for the week. It was a good week.
Oh yeah - Airbnb and poutine are really good - although, not necessarily related.
Monday, January 19, 2015
A Quick Update
Recently I attended the American Astronomical Society’s 225th conference in Seattle, WA. Even though the conference is a week long, I was only in attendance for one day.
I had a great time on my one day in Seattle; it was very full, and several people stopped by to see my poster. I also had meetings with other astronomers from around the country.
One encounter stuck out while speaking with someone at my poster. Even though he and I had an interest in the same science, he had never heard of me, nor seen my work. I actually expect this when I encounter people. I am under no illusion that my papers stand out from the dozens that get posted on astro-ph every day. Furthermore, I only have a small few papers. So, I was not surprised when he did not know me. By chance, I recalled his name from having read his paper, so in that way it was a little lopsided.
In any case, my point is that he would never have seen my work if I had not met him at the conference. We had a good conversation, and because he seemed genuinely interested, I followed up with him by sending an email the week after the conference.
The week after the conference I was laid out with a cold. This was very frustrating for me, because I couldn’t focus on anything to get work done. I did manage to bring my wits about me enough to apply for a couple jobs, so that was productive.
Speaking of . . . I’m really hoping to get a good job offer this year, because I’m trying to solve a two-body problem. Even though the nature of post-doc-hood and academia isn’t particularly partner friendly, I have to say that the people I have met within astronomy have all been understanding of the issue and have tried to be flexible to accommodate my situation.
This week I plan to go visit a colleague at another university. I am hoping to brainstorm a good ALMA proposal to take advantage of my Chilean observer status. While I am there I will give a talk to a general science audience. I like these kinds of talks, because they tend to be concept-oriented. I am looking forward to the week, but putting together the talk will take a bit of time. Next week I will have to get back to my narrow-line seyfert 1 data, so I can have some good results for the upcoming conference in March!
Phew, catch ya later
I had a great time on my one day in Seattle; it was very full, and several people stopped by to see my poster. I also had meetings with other astronomers from around the country.
One encounter stuck out while speaking with someone at my poster. Even though he and I had an interest in the same science, he had never heard of me, nor seen my work. I actually expect this when I encounter people. I am under no illusion that my papers stand out from the dozens that get posted on astro-ph every day. Furthermore, I only have a small few papers. So, I was not surprised when he did not know me. By chance, I recalled his name from having read his paper, so in that way it was a little lopsided.
In any case, my point is that he would never have seen my work if I had not met him at the conference. We had a good conversation, and because he seemed genuinely interested, I followed up with him by sending an email the week after the conference.
The week after the conference I was laid out with a cold. This was very frustrating for me, because I couldn’t focus on anything to get work done. I did manage to bring my wits about me enough to apply for a couple jobs, so that was productive.
Speaking of . . . I’m really hoping to get a good job offer this year, because I’m trying to solve a two-body problem. Even though the nature of post-doc-hood and academia isn’t particularly partner friendly, I have to say that the people I have met within astronomy have all been understanding of the issue and have tried to be flexible to accommodate my situation.
This week I plan to go visit a colleague at another university. I am hoping to brainstorm a good ALMA proposal to take advantage of my Chilean observer status. While I am there I will give a talk to a general science audience. I like these kinds of talks, because they tend to be concept-oriented. I am looking forward to the week, but putting together the talk will take a bit of time. Next week I will have to get back to my narrow-line seyfert 1 data, so I can have some good results for the upcoming conference in March!
Phew, catch ya later
Subscribe to:
Posts (Atom)