Introduction to GANs

Discriminator learns from real dataset and make judgement of input data.
Generator uses noise vestor to generate fake data and feed into discriminator, based on the feedback it can generate better samples.

Planning

Planning means identifying appropriate actions and their sequence, it’s the basis of rational behavior.
We need to design intelligent machines because:

1. the machine may be out of reach, eg. satellites
2. experts are not always available.

So the machine should learn how to work rationally by itself.

Introduction

Alpha Go: beat the best human Go player
AlphaFold: a solution to the protein folding

Reinforcement Learning is a feedback-based machine learning technique: an agent learns to behave in an environment by performing the actions and getting the feedback.
Sequential decision making + long-term goal(getting the maximum positive rewards)
No pre-program and only learns from its own experience

Agents

Perceives its environment through sensors and acts upon the environment through actuators
Rational agent: Acts to achieve the best (expected) outcome
For each possible percept sequence, a rational agent should select an action that is expected to maximize its performance measure, given the evidence provided by the percept sequence and whatever built-in knowledge the agent has.

Introduction

Regular ML: build a model based on sample data $x$ in order to make prediction
Attack: add invisible noise $\eta$ to clean sample data and fool the model(make the wrong prediction)
Adversarial Machine Learning: attempts to fool models by supplying deceptive input
The malfunction of a ML model may cause damage in many aspects.

java里hashmap是如何实现的？

JDK1.7：数组+链表
JDK1.8：数组+链表+红黑树

HashMap 为什么线程不安全？为什么ConcurrHashMap线程安全？

jdk 1.7使用的是头插法，在多线程环境下，使用HashMap的put操作会引起死循环，原因是多线程会导致HashMap的Entry链表形成环形数据结构，导致Entry的next节点永远不为空，就会产生死循环获取Entry。
ConcurrentHashMap 采用分段锁的方式，即每一段数据配一把锁，当一个线程占用其中一段时，其它段的数据仍然可以put或get。

进程和线程

进程与线程的区别

进程是资源（CPU、内存等）分配的基本单位，它是程序执行时的一个实例。程序运行时系统就会创建一个进程，并为它分配资源，然后把该进程放入进程就绪队列，进程调度器选中它的时候就会为它分配CPU时间，程序开始真正运行。
线程是程序执行时的最小单位，它是进程的一个执行流，是CPU调度和分派的基本单位，一个进程可以由很多个线程组成，线程间共享进程的所有资源，每个线程有自己的堆栈和局部变量。线程由CPU独立调度执行，在多CPU环境下就允许多个线程同时运行。同样多线程也可以实现并发操作，每个请求分配一个线程来处理。

Paper: https://arxiv.org/abs/1902.04885

Introduction

As the technology grows rapidly, more and more attentions are paid to Artificial Intelligence, one of the most famous examples is Alpha Go. It used 30000 games to train the model and finally defeated the top human Go players. The current rapid development of AI largely depends on the availability of Big Data.

Paper: https://arxiv.org/abs/1902.04885

Name: HUANG LiuEmail: e0575772@u.nus.eduStudent ID: A0225138J e_step(x_list, pi, A, phi)Initialize arrays alpha and c with zeros. For each sequence($x_list[o]$) in $x_list$: $alpha[o][0][k] = pi[k] p ...