Comments from IBM China Research Institute offer-capability is an attitude

When I reported the technical report to several interviewers at the IBM China Research Institute in Beijing over 30 minutes on the remote screen, I felt uneasy ...... It's not easy, but it's amazing!

In the next few days, I was very pleased to receive a call from two senior managers from IBM, who respectively introduced me to their respective departments and projects, indicating that my report was "impressive ", "outstanding capabilities "...... Thank you very much for giving me this opportunity!

Tips 

I am not a brilliant genius, but I believe in my research ability. This comes from a trick-I have come up with a "law", that is: ability is an attitude!

A simple explanation is as follows: there is no lack of smart people in this world, but there is a lack of people who know how to use their own talents.

Today's post, I hope to interpret this law through seven real stories-ability is an attitude!

1. I conducted an experiment in the research group via QQ and sent a group of six members to the same question. Three of them replied to me: "I have never encountered this problem, so I won't do it ......"; Two of them went to Google and probably told me how to get them. One of them didProgramI tested the advantages and disadvantages of different methods and told me what the best method is. After many years, many problems have been reported as "never met, no ......" People who have accumulated abilities are as smart as they are, but the gap is huge.

2. At the early stage of pedestrian simulation system development, I suddenly found *AlgorithmThis is particularly interesting with the social force model. It took some time to thoroughly study them... After I happily introduced various model algorithms to my boss, the boss said, "It is useless to introduce these model algorithms to the leaders. What they want is the effect, the model is endless ...". I was frustrated at the time, but it turned out that my "out of service" was right! Intelligent Dynamic pathfinding has become the core innovation of our system and the most proud content of each project introduction.

3. the "auxiliary decision-making system for urban rail transit operations" for Beijing was to be launched before the 08 Olympics. The time was very short. when the system was being debugged, the Beijing tester suddenly called and said that a candidate path between some stations seemed to be missing ...... At that time, everyone thought it could be a boundary condition problem. Just change it a bit. I have studied this K short circuit algorithm (developed by others) and found that it is a theoretical defect ......
At half past one, I couldn't explain it to the leaders clearly. I decided to rewrite this part and use the data to explain it. Due to the time constraints, I read many documents on the train back to Shanghai in Beijing. With a good a * algorithm Foundation, I quickly designed new algorithms. Tests show that the old algorithm has lost more than 500 paths! (Tens of thousands or so in total) at this time, everyone finally breathed a sigh of relief ......
But I didn't give up, because the algorithm speed was not fast due to the rush. It took several days to optimize the computing time for the division of more than 20 thousand OD in Beijing rail network to more than 10 minutes, and finally to 1 minute (in my notebook ). On the day of online debugging, the leaders praised: "This algorithm is really fast and accurate !"

4. The story of the above system: At that time, the basic data of Beijing Road Network was input by a master's degree. after his graduation, no data of Shanghai Road Network was obtained, and my boss asked me to do it. Although it's only half a day of physical activity, it's not a taste in my heart ......
Although someone persuaded me: "it takes a long time to get it done !" However, I am determined not to be stupid-It took me a week to design an intelligent Basic Data Management Subsystem Based on the "two-dimensional vector graphics library" that I developed, enter the station name andCode, The system automatically identifies, and then automatically generates all the data required for six database tables, including the range and transfer relationship. Later, the research team used this tool to build many road networks. Because it was very simple, this subsystem became the basis of the network Passenger Flow Simulation System in 863.

5. The logistics system class teacher mentioned a famous NP problem-Vehicle Routing Problem, asking everyone to go back and write a System Design book. At that time, I decided to develop this system. In the next few weeks, I found that the laws of genetic algorithms and nature were so consistent that they reached the crazy level. My girlfriend laughed at the following: "staying in the house all day ..." Later, the result was that the genetic algorithm I designed not only solved the problem that requires at least many vehicles, but also found a solution with a short total mileage.

6. The visit to Stanford is mainly involved in the study of an evacuation simulation system. However, due to the genetic algorithm background, another professor introduced me to one of their subjects-the auxiliary decision-making system for the office building transformation and optimization solution.
At the beginning, I thought this was a deterministic problem. Therefore, the use of the * algorithm has achieved better results and can meet the project requirements. I designed a genetic algorithm for comparison and found that in a few cases, I could "mutate" a better solution. After a large number of experiments, I found that although the two algorithms have different principles, there is an internal link in the data structure, and they can be combined into a universal framework to solve a large number of discrete optimization problems.
Out of my obsession with science, I did not stop-I designed a Data Structure to replace the hash table, the performance of the two algorithms was increased by 10 times at the same time (the concept of "roster" was proposed in the genetic algorithm), and a "crossover algorithm" was developed later ", the genetic algorithm is improved by more than 10 times. At that time, the person in the test case said that he could not keep up (he had already done it many times), so in the end, the data in paper was not obtained by my fastest algorithm.

7. Using the above "Discrete Optimization Problem search framework", I found that it can also solve many problems in the beauty of programming. When everyone is busy looking for a job for an interview, I spent a whole month studying the beauty of programming in my bedroom. Sometimes it takes a whole day to challenge a question, at that time, people around me said that I was "out of service" and I had some doubts. However, facts prove that this research not only proves interest, but also proves my algorithm capabilities, which is very helpful for finding a job later.

Conclusion 

Through the story above, I want to prove my law-ability is an attitude. If you want to ask what attitude is, I think it is a pure scientific attitude without any utilitarian attitude and persistent in the problem itself.

These seven stories can also be serialized using the figure below. I printed this picture for a short time and used it to impress facial officers.
 

In the study of the beauty of programming, you have seen many good ideas. However, they all come up when they walk, eat, and walk. I am never good at exams, and my girlfriend laughs at me. If you go to Microsoft for an interview, it may be dangerous, because you don't have time to walk ......

I once thought that for a doctor, the interview may not necessarily show his level, but if I give him 30 minutes of PPT time to introduce his research, the level will be clear at a glance.

What I did not expect was that the last two rounds of interviews with IBM all confirmed my previous ideas ...... Using PPT to give a speech is what I do best, because it has been honed too many times :)

Microsoft's technology is something I have studied and loved for many years. Microsoft is also a company I have dreamed of for many years. But now, I have hesitated. IBM's White Paper "the smart Earth wins in China" is indeed shocking-it is able to consider China's future from the perspective of the world's economic environment ......