1. From Bits to Qubits

Bits and Classical Computers

But, in computer science:

• A bit is the smallest (classical) unit of information that must be either 0 or 1.
• Bits can represent any piece of information. For example, we can represent the number \(2730_{10}\) in binary as \(101010101010_2\) , the character H as \(1001000_2\) , the symbol @ as \(1000000_2\) . More example here
• If you have \(n\) bits, they can be in \(2^n\) different classical states.

What is a (classical) computer?

• A computer is an electronic device for storing and processing data, typically in binary form (bits), according to instructions given to it in a variable program.

Qubits and Quantum Computers

What is a qubit?

• A qubit or quantum bit is the basic unit of quantum information, an extension of the bit to quantum mechanics.
• Qubit is a quantum variant of the bit, which means a qubit has exactly the same restrictions as a normal bit: it can store a single piece of information and can only give the output of 0 or 1. However, qubits can also be manipulated in ways that can only be described by quantum mechanics.
• It is a two-state quantum-mechanical system and one of the simplest systems displaying the peculiarity (essential characteristic) of quantum mechanics.
• A quantum state can be in superposition (i.e. simultaneously in 0 and 1). Superposition allows calculations to be performed on many states at the same time. So, some quantum algorithms have exponential speed-up.

What is a quantum computer?

Similar to a standard digital computer, which represents all information in the form of binary strings (or classical bits), the main difference between quantum computers and classical computers is that they use qubits instead of bits to represent information.

What is a quantum simulator?

A quantum simulator is a standard computer that calculates what an ideal quantum computer would do. Simulations are only possible for small numbers of qubits (~50 qubits). An example is the QASM Simulator by IBM Quantum.

2. Quantum Circuits at a glance

Computation as a Circuit

• Computation is the manipulation process to turn the inputs we have into the outputs we need (either using bits or qubits)
• A circuit diagram is a representation of the computation process.

Example: Two following bit-based and qubit-based circuit diagrams are equivalent

Bit-based circuit with AND and OR gates

Qubit-based circuit with NOT, CNOT, and Toffoli gates

What is a Quantum circuit?

In a quantum circuit, we typically need to do 3 jobs:

1. Encode the Input
2. Do some actual computation on the input
3. Extract an output

Example: a First quantum circuit with Qiskit

n = 8
qc = QuantumCircuit(n)
qc.x(7)
for j in range(n):
qc.measure(j,j)
qc.draw()

1. Encoding an input: Using NOT gate to flip final qubit (q7) from 0 to 1: \( 10000000 = q_7q_6q_5q_4q_3q_2q_1q_0\) (right to left)
   Note: Qiskit numbers the bits from right to left. Why?
   It has advantages when using bits to represent numbers. Qubit \(n^{th}\) is telling how many \(2^n\)s we have in the number. For example: \( 10000000 = 2^7 = 128 \)
2. Computation: Do nothing, keep the input unchanged
3. Output: Extract the output by using measurement - 1000000

3. Example: Build our first Quantum Circuit to do the addition

What is the Half Adder?

The Half Adder is a type of combinational logic circuit that adds two of the 1-bit binary digits. It generates the carry and sum of both the inputs.
– The Half Adder does not add the carry obtained from the previous addition to the next one.– The logic circuits of Haft-Adder and Full-Adder– What is about the Full Adder?
    

The Full Adder is also a type of combinational logic that adds three of the 1-bit binary digits for performing an addition operation. It generates a sum of all three inputs along with a carrying value.– The Full Adder, along with its current inputs A and B, also adds the previous carry.

Key ideas of the Half Adder

• Four basic sums:
  0+0 = 00 (in decimal, this is 0+0=0)
  0+1 = 01 (in decimal, this is 0+1=1)
  1+0 = 01 (in decimal, this is 1+0=1)
  1+1 = 10 (in decimal, this is 1+1=2)
• Key ideas of Half Adder:
  – The rightmost bit is completely determined by whether the two bits we are adding are the same or different.Specifically, if the two bits are equal (i.e., 0 + 0 or 1+1), the rightmost bit of the answer comes out 0 and otherwise.
  → Solution: Use the XOR gate (or CNOT gate in quantum computation)
  – The leftmost bit is only 1 if both of the bits we are adding are 1
      → Solution: Using the AND gate (or Toffoli gate in quantum computation)– We don’t want to overwrite the input
      → Solution: Using a different pair of qubits to write the output

What are an XOR and a CNOT gate?

• An XOR gate implements an exclusive or; that is, a 1 output results if one, and only one, of the inputs to the gate is 1. If both inputs are 1 or both are 0, the output result is 0.

A CNOT (Controlled-NOT) gate is a (reversible) quantum version of an XOR gate

– CNOT gate applies to a pair of qubits. One acts as the Control qubit (with the little dot, Input A), and another acts as the target qubit (with the big circle and +, Input B)– If the control qubit is 1, the CNOT gate does a NOT (flip) on the target qubit. Otherwise, it does nothing.

What are an AND and a Tofolli gate?

• An AND gate implements logical conjunction (∧) from mathematical logic. If (and only if) both inputs are 1, the output result is 1; otherwise, the output is 0
• The Toffoli gate, invented by Tommaso Toffoli, is a universal reversible logic gate, which means that any reversible circuit can be constructed from Toffoli gates. It is also known as the controlled-controlled-not gate, which describes its action. It has 3-bit inputs and outputs; if the first two bits are both set to 1, it inverts the third bit, otherwise, all bits stay the same.
• The Toffoli gate is essentially the atom of mathematics. It is the simplest element, from which every other problem-solving technique can be compiled.

Implement the Quantum Half Adder circuit

Example: The input is 11 (perform the 1 + 1 addition)

Qiskit code:

qc_ha = QuantumCircuit(4,2)
# encode inputs in qubits 0 and 1
qc_ha.x(0) # For a=0, remove the this line. For a=1, leave it.
qc_ha.x(1) # For b=0, remove the this line. For b=1, leave it.
qc_ha.barrier()
# use cnots to write the XOR of the inputs on qubit 2
qc_ha.cx(0,2)
qc_ha.cx(1,2)
# use ccx to write the AND of the inputs on qubit 3
qc_ha.ccx(0,1,3)
qc_ha.barrier()
# extract outputs
qc_ha.measure(2,0) # extract XOR value
qc_ha.measure(3,1) # extract AND value

qc_ha.draw()

Simulate and get the result (Qiskit <1.x):

counts = execute(qc_ha,Aer.get_backend('qasm_simulator')).result().get_counts()
plot_histogram(counts)

Result: 1 + 1 = 10 (binary). Or 1 + 1 = 2 (decimal)

References

1 Qiskit Online Textbook - Quantum States and Qubits by IBM Quantum (https://learn.qiskit.org/course/ch-states)

2 IBM Quantum Summer School 2021 (https://learn.qiskit.org/summer-school/2021)

2 Quantum Half-adder and Full-adder by Lahiru Madushanka (https://lahirumadushankablog.wordpress.com/2020/02/04/quantum-half-adder-and-full-adder/)

Ready to deep dive into the Federated Learning journey with selected state-of-the-art and valuable readings!

🎯 Intro

Blog

1. Google – Federated Learning: Collaborative Machine Learning without Centralized Training Data – April 6, 2017
2. Comic – https://federated.withgoogle.com/
3. https://medium.com/@ODSC/what-is-federated-learning-99c7fc9bc4f5
4. https://towardsdatascience.com/introduction-to-federated-learning-and-challenges-ea7e02f260ca
5. https://towardsdatascience.com/how-federated-learning-is-going-to-revolutionize-ai-6e0ab580420f

Initial papers

1. Practical Secure Aggregation for Privacy-Preserving Machine Learning – Google, 2016
2. Federated Learning: Strategies for Improving Communication Efficiency – 2016
3. Communication-Efficient Learning of Deep Networks from Decentralized Data – 2017

Talk – Seminar

1. Federated Learning One World Seminar – https://sites.google.com/view/one-world-seminar-series-flow/archive
2. Coursera – https://www.coursera.org/learn/advanced-deployment-scenarios-tensorflow

📜 Survey

1. Advances and Open Problems in Federated Learning
2. A Survey on Federated Learning Systems: Vision, Hype and Reality for Data Privacy and Protection
3. A Systematic Literature Review on Federated Machine Learning: From A Software Engineering Perspective
4. Federated Learning for 6G Communications: Challenges, Methods, and Future Directions
5. Threats to Federated Learning: A Survey
6. Federated Learning for Resource-Constrained IoT Devices: Panoramas and State-of-the-art
7. Federated Learning: Challenges, Methods, and Future Directions
8. Federated Learning in Mobile Edge Networks: A Comprehensive Survey
9. Federated Machine Learning: Concept and Applications
10. Towards Utilizing Unlabeled Data in Federated Learning: A Survey and Prospective
11. A Review of Privacy-preserving Federated Learning for the Internet-of-Things

📦 System design – frameworks – libraries

1. PySyft – A library for computing on data you do not own and cannot see
   • Github: https://github.com/OpenMined/PySyft
2. Tensorflow Federated
   • Website: https://www.tensorflow.org/federated
3. FedML: A Research Library and Benchmark for Federated Machine Learning
   • Website: https://fedml.ai/
4. Flower – A Friendly Federated Learning Framework
   • Github: https://github.com/adap/flower
5. Federated Learning Pytorch
   • Github: https://github.com/AshwinRJ/Federated-Learning-PyTorch
6. PrivacyFL: A simulator for privacy-preserving and secure federated learning.
7. Towards Federated Learning at Scale: System Design

💻 Models and Applications

1. DIOT: A Federated Self-learning Anomaly Detection System for IoT
2. MD-GAN: Multi-Discriminator Generative Adversarial Networks for Distributed Datasets.
3. (GAN) Federated Generative Adversarial Learning
4. Efficient Privacy-Preserving Edge Computing Framework for Image Classification

🛡️ Security and Privacy

Overview

An Overview of Federated Deep Learning Privacy Attacks and Defensive Strategies. 2020-04-01 Citation: 0

Backdoor Attacks

Awesome Backdoor Learning List: https://github.com/THUYimingLi/backdoor-learning-resources
Survey: Backdoor Learning – a survey

1. AISTATS 2020 – How To Backdoor Federated Learning ✅ ⭐️⭐️
   • Github: https://github.com/ebagdasa/backdoor_federated_learning
   • Conf: http://proceedings.mlr.press/v108/bagdasaryan20a.html
   • Supplementary PDF: http://proceedings.mlr.press/v108/bagdasaryan20a/bagdasaryan20a-supp.pdf
2. Blind Backdoors in Deep Learning Models (2021) ✅ ⭐️⭐️
   • Github: https://github.com/ebagdasa/backdoors101
3. ICLR 2020 – DBA: Distributed Backdoor Attacks against Federated Learning
   • Github: https://github.com/AI-secure/DBA
4. NeurIPS 2019 – Can You Really Backdoor Federated Learning?
5. Attack of the Tails: Yes, You Really Can Backdoor Federated Learning
6. NeurIPS 2020 Submission: Backdoor Attacks on Federated Meta-Learning

Data Poisoning

1. Data Poisoning Attacks Against Federated Learning Systems ✅ ⭐️⭐️
   • Github: https://github.com/git-disl/DataPoisoning_FL
2. Data Poisoning Attacks on Federated Machine Learning
3. Poisoning Attacks with Generative Adversarial Nets
4. Poisoning Attack in Federated Learning using Generative Adversarial Nets

Model Poisoning

1. ICML 2019 – Analyzing Federated Learning through an Adversarial Lens ✅ ⭐️⭐️⭐️ – Citation: 165 – Highlight: client attack
2. USS 2020 – Local Model Poisoning Attacks to Byzantine-Robust Federated Learning

Inference Attacks

1. Exploiting Defenses against GAN-Based Feature Inference Attacks in Federated Learning

Free-rider Attacks

1. NeurIPS 2020 – Free-rider Attacks on Model Aggregation in Federated Learning
2. Free-riders in Federated Learning: Attacks and Defenses

Leakage

1. Deep Models Under the GAN: Information Leakage from Collaborative Deep Learning. ACM CCS 2017. 2017-02-14 Citation: 284
2. Beyond Inferring Class Representatives: User-Level Privacy Leakage From Federated Learning. INFOCOM 2019 Citation: 56 Highlight: server-side attack
3. A Framework for Evaluating Gradient Leakage Attacks in Federated Learning. 2020-04-22 Researcher: Wenqi Wei, Ling Liu, GaTech
4. Quantification of the Leakage in Federated Learning. 2019-10-12 Citation: 1

Privacy

1. Privacy is What We Care About: Experimental Investigation of Federated Learning on Edge Devices
2. Comprehensive Privacy Analysis of Deep Learning: Passive and Active White-box Inference Attacks against Centralized and Federated Learning. 2018-12-03 Citation: 46
3. Inverting Gradients – How easy is it to break privacy in federated learning? 2020-03-31 Citation: 3
   • Github: https://github.com/JonasGeiping/invertinggradients

Defense

1. Mitigating Sybils in Federated Learning Poisoning. 2018-08-14. RAID 2020 Citation: 41 Highlight: defense
   • Github: https://github.com/DistributedML/FoolsGold
2. RSA: Byzantine-Robust Stochastic Aggregation Methods for Distributed Learning from Heterogeneous Datasets, AAAI 2019 Citation: 34
   • Github: https://github.com/Liepill/RSA-Byzantine
3. Towards Realistic Byzantine-Robust Federated Learning. 2020-04-10
4. FDA3 : Federated Defense Against Adversarial Attacks for Cloud-Based IIoT Applications. 2020-06-28
5. Learning to Detect Malicious Clients for Robust Federated Learning. 2020-02-01
6. Robust Aggregation for Federated Learning. 2019-12-31 Citation: 9
7. Towards Deep Federated Defenses Against Malware in Cloud Ecosystems. 2019-12-27
8. Attack-Resistant Federated Learning with Residual-based Reweighting. 2019-12-23
9. Robust Federated Learning with Noisy Communication. 2019-11-01 Citation:
10. Abnormal Client Behavior Detection in Federated Learning. 2019-10-22 Citation: 3
11. Byzantine-Robust Federated Machine Learning through Adaptive Model Averaging. 2019-09-11
12. An End-to-End Encrypted Neural Network for Gradient Updates Transmission in Federated Learning. 2019-08-22
13. Ensemble Distillation for Robust Model Fusion in Federated Learning
    • https://github.com/epfml/federated-learning-public-code/tree/master/codes/FedDF-code

Other Resources and References

• Awesome Federated Learning: https://github.com/chaoyanghe/Awesome-Federated-Learning (main reference)
• Google AI: https://ai.google/
• Deep AI: https://deepai.org/
• Papers with Codes: https://paperswithcode.com/task/federated-learning

In this article, we will set up a serverless platform with OpenFaaS, an open-source framework, to explore this paradigm—the first step in a deep dive into the serverless journey.

Prerequisites

1. One or 2+ servers (recommended), using Ubuntu 18.04.3 (Desktop or Server version, local or remote). Note: These servers should have a fixed IP address.
2. Internet connection

Setup Kubernetes Cluster with Rancher 2.0

There are manifold options for deploying a local or remote Kubernetes cluster. You can find out them here. One of the easiest ways to build one is Rancher.

Rancher is known as a platform to deliver Kubernetes-as-a-Service across any infrastructure and it could be used in both production and development environment.

In this experiment, I used 3 virtual machines running Ubuntu 18.04. All of them used sharing networking, so they could connect to each other and have internet access. At least one plays a master role (Master node), and others are Worker nodes.

Step 1: Install the latest version of Docker

Docker is a mandatory ingredient and must be installed in all nodes (both Master and Worker)

Option 1: Using the following script

curl -fsSL https://get.docker.com -o get-docker.sh
sh get-docker.sh

Option 2: Do it step-by-step

Install docker:

sudo apt install -y docker.io

After the installation is complete, start the docker service and enable it to run it every time at system boot.

systemctl start docker 
systemctl enable docker

Grant privileges to manage Docker to non-root user (don’t have to use “sudo docker”)

sudo groupadd docker 
sudo usermod -aG docker $USER

You can check it using the commands below:

systemctl status docker
docker —version

Step 2: Deploy the Rancher Server on the Master node

Login to the Master node and install Rancher. It is nothing more than running a container. Following the official docs, from your shell, enter the following command:

sudo docker run -d --restart=unless-stopped -p 80:80 -p 443:443 rancher/rancher

It will take some minutes to complete. Once it’s done, log in to Rancher by the IP https://<server_ip> (replace <serverip> with your Master node IP)

Then, you will be asked to set a password for the admin account and confirm the Rancher Server URL

Step 3: Setup the Kubernetes (K8s) Cluster

After login into the Rancher dashboard, click “Add Cluster” on the upper right corner. Then, select From existing nodes (Custom) to continue.

Type Cluster name, keep other fields as default or customize them if needed

You will see the command to deploy Master and Worker nodes afterward:

There are two cases: single-node cluster and multi nodes cluster

Single Node Cluster:

If you only have one server play the role of both Master and Worker:

• Check all boxes in part 1 – Node Options (etcd, Control Plan, Worker)
• Then, copy the command in part 2 and run it on your shell on your server. The command will be looked like:

sudo docker run -d --privileged --restart=unless-stopped --net=host -v /etc/kubernetes:/etc/kubernetes -v /var/run:/var/run rancher/rancher-agent:v2.3.2 --server https://<server_ip> --token wxhkwshb69zjk76jfrgmw8z8raabbccf98xjkv4fh92gmcrcm6d26r --ca-checksum cc2ccdb2e5a2620a4daec8de9fb100daabbccad85d8c169a1b1ddaa6fe6895d5 --etcd --controlplane --worker

Multi Nodes Cluster:

If you have 2 nodes or more than, then one is Master and others are Worker. You should set up follow the following rules:

• Master node: check etcd and Control Plan and use the command in part 2 to run on Master node
• Worker node: check the only Worker and use the command in part 2 to run on all Worker nodes.

Note that Rancher uses the hostname of each server as the Node name by default. If you have 2 or more nodes with the same hostname, you should define the different Node names for each other. Otherwise, the latter nodes couldn’t be added to the Cluster.

After that, wait for the provisioning process (it could take some minutes). When it’s done, you can see all nodes are activated in the Nodes page

Step 4: Setup OpenFaaS – Serverless platform on Rancher K8s Cluster

OpenFaaS is a prevalent open-source platform for Serverless (besides OpenWhisk, Kubeless, Fission,…). OpenFaaS is Kubernetes-native and uses Deployments, Services, and Secrets. For more detail check out the “faas-netes” repository. Presently, OpenFaaS is not available in the library of Rancher so we will need to install it by hand. Once you have a cluster, you can follow these steps. I will deploy by kubectl, you can try other options (like using Helm or k3sup by find out on this page)

Install kubectl

Login to the Master node and run the following command on your shell

sudo snap install kubectl --classic

Put the Rancher Kubeconfig file into the Master node

Login to Kubernetes dashboard, on Cluster page, click Kubeconfig File on the upper right corner. You will see the kubeconfig file in the new popup.

Then copy all and put this into the file in /home/<user-name>/kube/config (create a new file if it does not exist).

Note that, you should change the permission of this file to ensure it could be accessible by the following command: sudo chmod 755 /home//.kube/config

Install OpenFaaS CLI

You can install the OpenFaaS CLI (to interact with CLI) using brew or a curl script.

Via curl:

curl -sL https://cli.openfaas.com | sudo sh

Via brew:

brew install faas-cli

Deploy OpenFaaS from faas-netes YAML and kubectl

Clone the repository:

git clone https://github.com/openfaas/faas-netes

Deploy the whole stack: This command is split into two parts so that the OpenFaaS namespaces are always created first:

• openfaas – for OpenFaaS services
• openfaas-fn – for functions

Enter to faas-netes folder and deploy openfass and openfaas-fn namespace

cd faas-netes
kubectl apply -f namespaces.yml --kubeconfig=/home/<user-name>/.kube/config

Create a password for the gateway:

# Generate a random password
PASSWORD=$(head -c 12 /dev/urandom | shasum| cut -d' ' -f1)
kubectl -n openfaas create secret generic basic-auth \
--from-literal=basic-auth-user=admin \
--from-literal=basic-auth-password="$PASSWORD"

Note: You can use your own password and remember it to login OpenFaaS UI. If using above command to generate a random password, then review and keep it (using echo $PASSWORD)

Deploy OpenFaaS:

In faas-netes folder, run following command:

kubectl apply -f ./yaml --kubeconfig=/home/<user-name>/.kube/config

Add OpenFaaS Project to Rancher

Back to Rancher dashboard, navigate to Projects/Namespace, you will see 2 namespaces openfaas and openfaas-fn are created. Then, click Add Project on the upper right corner, enter Project Name (like openfaas or other up to you) and click Create. Next, select 2 openfaas namespaces and click Move button (on upper left corner) to move them to Project openfaas you have just created. Result:

When clicking on your project, you will see all your Workloads

Set up OpenFaaS gateway to access OpenFaaS UI

Select Load balancing section, then click Add Ingress on the upper right corner. Type Ingress name, select Target is gateway workload, port like 8080, keep others fields á default and click Save.

The result will be like the following:

Config Gateway Environment

export OPENFAAS_URL=gateway.openfaas.<your_gateway_IP>.xip.io

Now, you can access the OpenFaaS URL via the URL gateway.openfaas.<server-ip>.xip.io (in Targets) section and login with username admin and password in step 3

Congratulation! OpenFaaS deployment on Rancher is completed and you can play with Serverless now.

References

https://github.com/openfaas/faas

Get started

Get a feel for quantum computing with interactive introduction:

1. Quantum Computing Journey by QuantumAI.Google
   An interactive map to get started with Quantum Computing from Google QuantumAI
2. Introduction to Quantum Computing by Qiskit Learn
   An short introduction course to get started with Quantum Computing
3. Quantum Country by Andy Matuschak and Michael Nielsen
   An introduction to quantum computing and quantum mechanics

Courses

1. Quantum Information Science - MAS.865J by Isaac Chuang and Peter Shor
   MIT Open Courseware
2. Introduction to Quantum Computing and Quantum Hardware by IBM Qiskit
   2020 Qiskit Global Summer School
3. Quantum Machine Learning by IBM Qiskit
   2021 Qiskit Global Summer School
4. Quantum Computing by Brilliant
   Solve hard problems by computing with quantum mechanics.
5. Quantum Computing A-Z by Syed-Mohammad Raza
   Learn Quantum Algorithms, Quantum Machine Learning, Quantum Information Theory, Quantum Error Correction & More!

Books

1. Quantum Computation and Quantum Information by Michael A. Nielsen & Isaac L. Chuang
   Comprehensive textbook and “bible” for Quantum Computing
2. Quantum Computing for Computer Scientists by Noson S. Yanofsky and Mirco A. Mannucci
3. Quantum Computer Science: An Introduction by N. David Mermin
4. Quantum Computing for the Quantum Curious by Ciaran Hughes, Joshua Isaacson, Anastasia Perry, Ranbel F. Sun, Jessica Turner
5. Qiskit Online Textbook by IBM Qiskit

Programming Languages and SDKs

1. Qiskit by IBM
   An open-source SDK for working with quantum computers at the level of pulses, circuits, and application modules.
2. Cirq by Google
   An open source framework for programming quantum computers
3. Q# (Q Sharp) by Microsoft
   A Microsoft’s open-source programming language for developing and running quantum algorithms
4. Braket SDK by Amazon
   An open source library that provides a framework that you can use to interact with quantum computing hardware devices through Amazon Braket.
5. PennyLane by Xanadu
   A cross-platform Python library for differentiable programming of quantum computers. Train a quantum computer the same way as a neural network.
6. PyQuil by Rigetti
   A Python library for quantum programming using Quil, the quantum instruction language developed at Rigetti Computing

Quantum Providers

1. IBM Quantum by IBM
2. Amazon Braket by Amazon
3. Azure Quantum by Microsoft
4. Google Quantum Computing Service by Google
5. Strangeworks QC™ by Strangeworks

Quantum Blogs

1. Quantum Landscape 2022 by Samuel Jaques, University of Oxford

To be updated

What is Quantum Computing?

Why do we need Quantum Computing?

When can we access a Quantum computer and develop a Quantum application? Where are we now in the Quantum world?

We will together address these questions in this blog post to help you get a feeling about the current state of quantum computing – a promising paradigm shift in the near future.

What is Quantum Computing?

Quantum computing bases on the theory of quantum mechanics, and therefore, is fundamentally different from classical computer.

Wait, what is quantum mechanics? Quantum mechanics is a branch of physics to actually explain the nature of phenomena at the scale of atoms.

But why fundamentally different?

We could start to answer this question from the basic unit of classical computer and quantum computer: bits and quantum bits (or qubits). As we have learned, the fundamental unit of information in today’s computer is the bit, which could be either 0 or 1 (we could call they are states). In other words, we could break every digital information into 0s and 1s. 1010 is the bit representation of 18, for example. However, quantum data are made up of qubits, which could be 0, 1 or a combination of 0 and 1 at the same time (we called a superposition state). That means, with 1 qubit, we could have 2 states simultaneously. How about if we have 3 qubits? — 8 states, right? 10 qubits? 1024 states at the same time with quantum computer while it still be 1 state once with classical solution.

Along with superposition, quantum computing have another important characteristic called entanglement. We can simply think about an example of entanglement is special situation in which 2 qubit are strong correlated regardless how far away between them. Strong correlated means that one qubit could always determine the state of the second qubit without interacting with the second one. Looks like we could transfer the quantum information faster than the speed of light using entanglement? It is really “spooky” (as Einstein dubbed).

Exploiting this characteristic, we could take some exponential speed-up when using quantum computing to solve some intractable issues of today’s computer. A bit fancy, but it’s not an easy game, there are many things we need to learn and control before reaching all the advantages. We will discuss more in later posts, don’t make it complicated at the beginning 🥲

Why do we need Quantum Computing?

For some problems, supercomputers aren’t that super.

Quantum computing is a promising technology to solve multiple intractable task even with “classical” supercomputers.

For example, factorising a large number into prime factors, which is the “Achilles heel” of crypto algorithms like RSA, could be “impossible” with the most powerful supercomputer. However, it’s not the rough way for quantum computers in near future by utilising Shor’s algorithm. This security risk has urged the emergence of another field called Post-Quantum Cryptography to develop a quantum-resistant crypto algorithms.

Apart from cryptography, quantum computing could be a solution for many prominent areas such as chemistry, finances, machine learning, networks and communications. That’s why now more and more countries put quantum computing on top of the national strategy to prioritise the large investment.

When can we access an actual Quantum computer?

Now! Why not?

Actually, from several years ago, we can access to actual quantum computer developed by IBM. The most popular way to run a program on quantum computer is through a cloud service, provided by some well-known vendors, such as IBM Quantum, Amazon Braket, Azure Quantum. That means, you can write your quantum application at your local computer and then send them to a quantum cloud service for executing, then waiting for the results. If you want to learn and try using quantum computer, go ahead to IBM Quantum - you can send your first quantum job to (up-to) 5-qubit quantum computers for FREE!

Where are we now in the Quantum world?

We are now staying in the NISQ – Noisy Intermediate-Scale Quantum era. That simply means the quantity and quality of qubits is not good enough to apply in practical problems at the moment. However, we witnessed numerous advancements in developing quantum hardware recently. The most powerful quantum computer presently (April 2022) is released by IBM last year with 127 qubits – a big milestone for the future of quantum computing. They also targeted to develop a 1121-qubit quantum computer by the end of next year, 2023. In 2019, Google also claimed that their quantum computer just need 200 seconds to solve a 10,000-year problem of the world’s fastest supercomputer. They called it “quantum supremacy”. Moreover, many research topics related to quantum computing becoming a trending and very “hot” topics recently, such as Quantum Machine Learning, Quantum Error Correction.

How can we get started to learn Quantum Computing?

When I started learning Quantum computing, I found this interesting quote from Frank Zickert (PyQML):

Quantum computing is fundamentally different from classical computing.
To master quantum computing, you must unlearn what you have learned.

But don’t worry too much! Nowadays, we can learn almost any of new things by ourselves by many ways. There are many resources to learn about quantum computing, and it depends on your curiosity, how deep you want to dive in and how you intend to use quantum for your work.

• If you would like to get a bit more “feeling” about quantum computing and quantum mechanics, check out a website calledQuantum Country (by Andy Matuschak and Michael Nielsen).
• If you want to have solid understanding about quantum, read the quantum “bible” entitles Quantum Computation and Quantum Information by Isaac Chuang and Michael Nielsen.
• If you want an interactive textbook to learn Quantum computing in a more practical way and start to write your first quantum program using your Python skill, check out Qiskit Textbook.

Just named a few, but there are numerous good books, online courses, tutorial videos, documentation available. If you are really interested to learn, define your keywords and your Google-ing skill will definitely help.

Happy learning!