FAQ

1. do you have any recommendations for books on machine learning that I could read?

   This is a question I frequently am asked by students looking to transition into machine learning or data science. My own experience is that it's helpful, with a new field, to read the same topic described by different people, particularly people coming from different backgrounds who emphasize different aspects. In 2018, for machine learning, you might start with these:

   1. Elements of Statistical Learning is written by people coming from applied computational statistics as a background and has the added advantages of being nearly comprehensive, well-known to other students, and free. This is a good "hello world" for machine learning as viewed by applied computational statisticians.
   2. Chris Bishop's Pattern Recognition and Machine Learning" is almost as well known, at least within the field of generative modeling in machine learning. Bishop's educational background was in theoretical physics, so you might find it appealing if you are coming from that background as well.
   3. Kevin Murphy's book on machine learning is quite popular. Kevin is a computer scientist by background.
   4. I really enjoy Foundations of Machine Learning, though it doesn't seem to be as widely-read by students as the above 3 books. In particular I find it much more mathematical, which may just be the kind of book I like, with good appendices relating the "need-to-know" bounds and inequalities on which much of the PAC framework in machine learning is based. Also it's got good sections on reinforcement learning, which is under-discussed in most books on ML.

2. I am an undergraduate doing research in X. Do you have a short-term research project I could work on?
   1. what subjects are you interested in and
   2. what programming languages are you most comfortable in and
   3. what research have you done and
   4. has any of it been published?
   
   
3. I am an undergraduate doing research in X. Would you fund my PhD?
   
   that's great that you've done research.
   
   i encourage you to apply but usually don't answer questions about i would accept a student into my group until i have worked with them in person for several weeks or months and have some sense for whether their research strengths are commensurate with my research interests. Also, if the student is not funded by fellowship, i need to have some sense for whether resources are likely to be available to support the student for the next several years until completion of the PhD.
   
   i hope that you will apply for admission to Columbia university. I've taken students from several departments, including physics, BME, chemistry, neuroscience, electrical engineering, and of course my own department: the department of applied physics and applied mathematics. I've also had students work with me through the medical school's integrative program as part of the c2b2 track. i encourage you to apply to whichever programs you think have the most interesting or least uninteresting graduation requirements and whose subject areas are close to your research background.

   
   
   
   

4. I am an undergraduate at {university}. Will you host me for the summer?
   
   thanks for your email.
   
   owing to the finiteness of time however i can't make time commitments to students I've never met or worked with. unfortunately i get many emails such as yours, even many emails from well-ranked universities, and it's simply not possible for me to commit time to host students i don't know.
   
   
5. I'm interested in any opening for your summer projects
   
   thanks for your email.
   
   1. please send CV
   2. are you looking for $, course credit, or experience?
   3. when do you need an answer?
   4. generally i do research with someone for a month before committing financially.
   5. please send copies of any writeups or publications
   6. please send a letter of reference from someone who has supervised your research in the past.
      
   
   
6. Does Columbia have a job placement office?
   • here

7. I want to sign up for 3900 (supervised research). How many credits will you give me?
   
   If you want to take 3900 with me, we need to come to a contract, and this contract needs to be closed before the start of the semester. The contract will stipulate:
   • Who is the scientific advisor (if not me)
   • What is the deliverable (e.g., technical report, oral report)
   • Deadline for the deliverable
   • How often will you update me on progress of the research and in what form
   • How many points of credit will be awarded
   The scientific advisor, me, and you must all come to agreement on the contract, and the scientific advisor must agree to email me a review of your work by a deadline stated in the contract. Example contract: write an email like this

     Dear Professor Wiggins
     
     I propose to study ${TOPIC} under the scientific direction of ${PERSON}.  
     I will update you each ${TIME_INTERVAL} with an email between ${MIN} and ${MAX} words in length.  
     On or before ${DATE}, I will 
       {deliver a talk of no less than ${TIME} minutes, 
       hand in a formal writeup of my results of no less than ${LENGTH} pages}, 
     for ${NUM_CREDITS} credits.
     
     Please approve
     
     -${name}, ${UNI}
     

8. I am an applied math major and I want to substitute class X for class Y
   
   Thanks
   • what year are you?
   • which undergrad division?
   • what is the conflicting class?
   • for what is it required?