[論文速速讀]系列文章介紹
論文速速讀系列是從今年四月開始，我開始寫一些論文的中文讀書筆記，還記得第一篇是[論文速速讀]ReZero is All You Need: Fast Convergence at Large Depth，之後發現儘管自己已經陸續產出了幾篇文章，可是好像都沒正式的跟大家介紹這系列文章的由來xD
所以這篇文章就是來講講這系列文章到底是什麼，以及我會和會想寫這些文章。
由於在AI領域每年總是有一些非常重大的突破和應用，如果跟不上潮流很有可能就會錯失許多機會。例如，對NLP領域熟悉的話你一定聽過2013年的word2vec、2014年開始流行的attention、2018年的Bert…這些很有名的技術。
還記得Bert剛出的時候我好像剛進碩士實驗室，當時只知道這個技術屌打了當時一堆NLP的研究，但我想也想不到兩年後Bert已經造成如此大的影響力，一堆基於Bert的變形應用在各大領域上都取得了非常優異的結果。
因此，我想要藉由這系列的文章讓自己能夠更加快速的了解AI的新技術和研究，同時逼迫自己看論文xD
German farmers are given a premium depending on the conditions at their farmyard. Imagine the following simplied regulation: you know the size of each farmer’s farmyard in square meters and the number of animals living at it. We won’t make a difference between different animals, although this is far from reality. Moreover you have information about the degree the farmer uses environment-friendly equipment and practices, expressed in a single integer greater than zero. The amount of money a farmer receives can be calculated from these parameters as follows. First you need the space a single animal occupies at an average. This value (in square meters) is then multiplied by the parameter that stands for the farmer’s environment-friendliness, resulting in the premium a farmer is paid per animal he owns. To compute the nal premium of a farmer just multiply this premium per animal with the number of animals the farmer owns.
You have just moved from Waterloo to a big city. The people here speak an incomprehensible dialect of a foreign language. Fortunately, you have a dictionary to help you understand them.
You have bought a car in order to drive from Waterloo to a big city. The odometer on their car is broken, so you cannot measure distance. But the speedometer and cruise control both work, so the car can maintain a constant speed which can be adjusted from time to time in response to speed limits, traffic jams, and border queues. You have a stopwatch and note the elapsed time every time the speed changes. From time to time you wonder, “how far have I come?”.
To solve this problem you must write a program to run on your laptop computer in the passenger seat.
The game Minesweeper is played on an n by n grid. In this grid are hidden m mines, each at a distinct grid location. The player repeatedly touches grid positions. If a position with a mine is touched, the mine explodes and the player loses.
If a position not containing a mine is touched, an integer between 0 and 8 appears denoting the number of adjacent or diagonally adjacent grid positions that contain a mine. A sequence of moves in a partially played game is illustrated below.
Here, n is 8, m is 10, blank squares represent the integer 0, raised squares represent unplayed positions, and the gures resembling asterisks represent mines. The leftmost image represents the partially played game. From the rst image to the second, the player has played two moves, each time choosing a safe grid position. From the second image to the third, the player is not so lucky; he chooses a position with a mine and therefore loses. The player wins if he continues to make safe moves until only m unplayed positions remain; these must necessarily contain the mines.
Your job is to read the information for a partially played game and to print the corresponding board.
PS:2014年，從我部落格轉過來的
因為學測國文採計而非常吃虧的我 好險面試過程還不壞…
摁…不說廢話了 面試場所長這個樣子 是3對3的形式 總共20分鐘
| 一直微笑的教授1號 | 看不到臉的教授2號 | 沒什麼特徵的教授3號 |
|---|---|---|
| 我 | 面試1號 | 面試2號 |
中間大約隔了一大張桌子的距離 為什麼會說看不到臉的教授…因為他被電腦檔住了
而1號2號…雖然這樣說有點無禮 但在整個過程中他們真的帶給了我蠻大的歡樂..
因為高雄大學資工面試的資料真的太少了(我只找到兩筆..) 因此決定來造福一下學弟妹…
PS:2014年的內容，從我的部落格轉過來的
面試前學長會先叫六個進去一間很空曠又很陰暗的房間
裏頭有六張椅子 + 一個功率非常高的冷氣不斷的吹你… 之後一批三個進去另外一個房間 房見內有3個辦公事 裡面住著3個BOSS…
不是…是教授 每一間你有7分鐘的攻略時間 時間到了門外會有學長將你強制登出(敲門- -)
