Dear all,
I have a simple question, that's stated in the subject of this post: Do I have a chance in the game industry?
Before you can answer, let me tell you a little bit about myself and what I want to do. I have a french engineering master's degree in Signal Processing for digital communication, which I obtained in 2010. After my master's degree I had a PhD contract with Technicolor in France. The subject of the PhD was the multimodal analysis of video content, and I especially worked on events recognition in movies audio streams, such as the detection of gunshots, or screams. The methods heavily relied on machine learning, and I like to think I acquired a fair expertise in the subject. I worked particularly with support vector machines (SVM), Bayesian Networks (which I know the best, strucutre learning, parameter learning, inference, you name it), Gaussian Mixture Models, Kmeans, K-Nearest Neighbours (who doesn't know about this one? :-)), Hidden Markov Models (HMM) and I have knowledge about a few others, such as the well know Neural Networks. In the end, I worked on how to stabilise the representation of the data that we input the classifier algorithm. I defended my PhD in October last year, 2013.
After working for Technicolor, I was fed up the multimedia industry, especially the fact that we are researching new ways to sell people advertisement, or new ways to have the user NOT think. So I decided to try my luck in another field. I also wanted to try the public lab environment. So I am now a Post-doc in a lab that works on surgical process modelling, and information processing in the OR, in collaboration with the University Hospital of Rennes in France. I myself work on process synthesis, using Petri Nets and their derivatives (another technique to add to my machine learning knowledge :-)). However I found that public system, and the fact that I actually work alone once again, is not something for me. I have no stress, no sense of urgency, and a good part of my work is about making stuff that I don't like, in a field that I don't believe. Let met elaborate on the latter for a bit. I am working on surgical process modelling, so I have to work with surgeons on how to represent the surgical processes. The idea is to have examples of surgical processes for one type surgery, and try to derive a general process model from them. This part is actually quite interesting, if you don"t consider the fact that you have to recolt the data, validate them, filter them before you can use them, which means about 3-4 days of work for one example (We need lots to have meaningfull results, that's just for one surgery type. Change the surgery type and start over...). I was asked to try to use ontologies (which I know very little about) in the process learning approach, but from what I gathered about ontologies, they are less than usefull for what we need... In other words, I don't believe in those, at least in what we want to do with it. In french, I use a term that literally translates as "Intellectual Wanking" (Pardon my french). And don't get me started on the validation process. I also don't believe that what we do in research will work before the next 20 years, or at least with a robustness and generalisation properties that would be enough for me to actually consider selling them.
I am not saying this last job taught me nothing, on the contrary. I now know about process mining, and I have increased my skills in programming and program optimisation. But I want to try to apply my knowledge of machine learning, especially the way they are used, and my knowledge of programming to stuff that will come out in the next future (2-3 years max), and that I would actually be proud of. I also want to work in a team of more than two people, where one of them isn't only telling the other what to do.
On the other hand, I always felt that what my knowledge of learning algorithms could be applied to video games, but I knew getting in the industry was hard, so I did not even considered applying. As the year passed, I wander around on the internet, on pages about how to make video games, etc... And one day I jumped and bought GCC4 along with "Artificial Intelligence for Games" from Ian Millington et. al. I went through the first one which convinced me that I want to do this, and I am reading the second one, which definitely confirmed my conviction. I want to be a programmer, if possible in the video games industry, and cherry on the cake, an AI programmer. I am currently learning about design patterns using my wife's course material (she is a computer science engineer), and I made my first Tetris (which I had a lot of fun making :-)). I am currently developing a space shooter, a danmaku to be specific, with a friend who actually plays a lot of them, but doesn't know how to program, so that's quite interesting. We are using Unity, and when it is ready, I'll put it on the web for you to try. Then I plan of building a small rugby IA to apply my IA4G knowledge.
I hope I haven't bored you to death with this long post. If had the courage to reach the end, I would appreciate if you could give me a piece of mind on how relevant my CV would be for the game industry. I know I have to finish at the very least one or two games before, but I feel like I can sell my PhD as experience.
Thank you very much for reading me,
--DerickTP
PS: don't hesitate to kill every hope I have. Better to tell a harsh truth than leave someone expecting.
I have a simple question, that's stated in the subject of this post: Do I have a chance in the game industry?
Before you can answer, let me tell you a little bit about myself and what I want to do. I have a french engineering master's degree in Signal Processing for digital communication, which I obtained in 2010. After my master's degree I had a PhD contract with Technicolor in France. The subject of the PhD was the multimodal analysis of video content, and I especially worked on events recognition in movies audio streams, such as the detection of gunshots, or screams. The methods heavily relied on machine learning, and I like to think I acquired a fair expertise in the subject. I worked particularly with support vector machines (SVM), Bayesian Networks (which I know the best, strucutre learning, parameter learning, inference, you name it), Gaussian Mixture Models, Kmeans, K-Nearest Neighbours (who doesn't know about this one? :-)), Hidden Markov Models (HMM) and I have knowledge about a few others, such as the well know Neural Networks. In the end, I worked on how to stabilise the representation of the data that we input the classifier algorithm. I defended my PhD in October last year, 2013.
After working for Technicolor, I was fed up the multimedia industry, especially the fact that we are researching new ways to sell people advertisement, or new ways to have the user NOT think. So I decided to try my luck in another field. I also wanted to try the public lab environment. So I am now a Post-doc in a lab that works on surgical process modelling, and information processing in the OR, in collaboration with the University Hospital of Rennes in France. I myself work on process synthesis, using Petri Nets and their derivatives (another technique to add to my machine learning knowledge :-)). However I found that public system, and the fact that I actually work alone once again, is not something for me. I have no stress, no sense of urgency, and a good part of my work is about making stuff that I don't like, in a field that I don't believe. Let met elaborate on the latter for a bit. I am working on surgical process modelling, so I have to work with surgeons on how to represent the surgical processes. The idea is to have examples of surgical processes for one type surgery, and try to derive a general process model from them. This part is actually quite interesting, if you don"t consider the fact that you have to recolt the data, validate them, filter them before you can use them, which means about 3-4 days of work for one example (We need lots to have meaningfull results, that's just for one surgery type. Change the surgery type and start over...). I was asked to try to use ontologies (which I know very little about) in the process learning approach, but from what I gathered about ontologies, they are less than usefull for what we need... In other words, I don't believe in those, at least in what we want to do with it. In french, I use a term that literally translates as "Intellectual Wanking" (Pardon my french). And don't get me started on the validation process. I also don't believe that what we do in research will work before the next 20 years, or at least with a robustness and generalisation properties that would be enough for me to actually consider selling them.
I am not saying this last job taught me nothing, on the contrary. I now know about process mining, and I have increased my skills in programming and program optimisation. But I want to try to apply my knowledge of machine learning, especially the way they are used, and my knowledge of programming to stuff that will come out in the next future (2-3 years max), and that I would actually be proud of. I also want to work in a team of more than two people, where one of them isn't only telling the other what to do.
On the other hand, I always felt that what my knowledge of learning algorithms could be applied to video games, but I knew getting in the industry was hard, so I did not even considered applying. As the year passed, I wander around on the internet, on pages about how to make video games, etc... And one day I jumped and bought GCC4 along with "Artificial Intelligence for Games" from Ian Millington et. al. I went through the first one which convinced me that I want to do this, and I am reading the second one, which definitely confirmed my conviction. I want to be a programmer, if possible in the video games industry, and cherry on the cake, an AI programmer. I am currently learning about design patterns using my wife's course material (she is a computer science engineer), and I made my first Tetris (which I had a lot of fun making :-)). I am currently developing a space shooter, a danmaku to be specific, with a friend who actually plays a lot of them, but doesn't know how to program, so that's quite interesting. We are using Unity, and when it is ready, I'll put it on the web for you to try. Then I plan of building a small rugby IA to apply my IA4G knowledge.
I hope I haven't bored you to death with this long post. If had the courage to reach the end, I would appreciate if you could give me a piece of mind on how relevant my CV would be for the game industry. I know I have to finish at the very least one or two games before, but I feel like I can sell my PhD as experience.
Thank you very much for reading me,
--DerickTP
PS: don't hesitate to kill every hope I have. Better to tell a harsh truth than leave someone expecting.
The post was edited 2 times, last by DerickThePoney ().