Workshop Title:

Autoencoder Based Semantic Communications for Image Transmission on Error Prone Channels

Date:

May 2nd, 2023 (GMT+1)

Organizer:

Department of Computer and Information Sciences, University of Strathclyde

Keywords:

• Semantic Communications
• Autoencoders
• Machine Learning
• Wireless Communications
• Polar Codes

Workshop Chair:

Personal Bio:

Prof. Anil Fernando received the B.Sc. (Hons.) degree (First Class) in electronics and telecommunication engineering from the University of Moratuwa, Sri Lanka, in 1995, and the M.Sc. in Communications (Distinction) from the Asisan Institute of Technology, Bangkok, Thailand in 1997 and Ph.D. in Computer Science (Video Coding and Communications) from the University of Bristol, UK in 2001. He is a professor in Video Coding and Communications at the Department of Computer and Information Sciences, University of Strathclyde, UK. He leads the video coding and communication research team at Strathclyde. He has worked on major national and international multidisciplinary research projects and led most of them. He has published over 400 papers in international journals and conference proceedings and published a book on 3D video broadcasting. He has been working with all major EU broadcasters, BBC, and major European media companies/SMEs in the last decade in providing innovative media technologies for British and EU citizens. His main research interests are in Video coding and Communications, Machine Learning (ML) and Artificial Intelligence (AI), Semantic Communications, Signal Processing, Networking and Communications, Interactive Systems, Resource Optimizations in 6G, Distributed Technologies, Media Broadcasting and Quality of Experience (QoE).

Workshop Committee Members:

Prabath Samarathunga, Researcher, buddhi.samarathunga-liyanage-don@strath.ac.uk

Workshop Description:

Background:

The use of multimedia services for everyday activities such as pictures, streaming, broadcasting, teleconferencing, video-on-demand and peer-to-peer video sharing has undergone unprecedented growth in the recent years. It has been forecasted that in 2024 almost 95% of global data traffic will be image and video, mainly fueled by vast numbers of consumer communication devices being introduced to the market, coupled with the users' higher consumption of multimedia services. Crucially, the quality of the received media is of prime importance to users, as well as service providers, irrespective of where from and how the users are connected to the service. Since these users are increasingly mobile, providing the necessary capacity to handle this ever-increasing media traffic poses significant challenges for the future communications infrastructure, especially in mobile-wireless systems where spectrum capacity and handset resources (e.g., battery capacity) are limited. The capacity-efficiency challenge is also evident in the image/video coding and connect processing front. Even though the latest standards, such as 5G technologies increase peak data rates in the downlink to as much as 10 Gbps, and the emerging video coding standards such as Versatile Video Coding (VVC) improves the coding efficiency by approximately 50% compared to the predecessor and state-of-the-art H.265/HEVC, this will still not be sufficient in practical networks, where capacity is shared between multiple users for voice, image, data and video services. This is especially true for bandwidth-hungry and resource intensive video applications that will adopt the upcoming high resolution video formats, such as Ultra High Definition (UHD), Super High Definition (SHD), High Dynamic Range (HDR), 3600 videos, 6 Degree-of-Freedom (6DOF) videos contents and real-time interactive multimedia applications such as ACTION-TV for a superior visual experience over existing conventional formats and technologies. This workshop addresses an autoencoder based semantic communication system to address the above issues.

Goal / Rationale:

The conventional approach to communication involves transmitting the minimum number of bits with minimal errors between two points. This was based on Shannon's 1948 paper, which established channel capacity and showed that rates below capacity are possible without incurring an exponentially higher number of errors at the receiver side. However, this approach does not explicitly exploit the information about the source available at the transmitter side. Semantic communication is a paradigm that aims to address the second layer of communication, the semantic problem, by delivering the semantic meaning of the message instead of the exact form of the message. Currently, there is no unique transmission strategy for semantic communication, so a system must be designed in accordance with the current communication framework. The challenge is to ensure that the transmitter's semantic information is preserved at the receiver while transmitting through the physical channel, and research is required for different media types over conventional communication standards with semantic communications. This research aims to develop an autoencoder based semantic communication system to transmit images over a channel to optimize the bandwidth while maintaining the quality of the image. A multi layered autoencoder is used to generate a latent vector for a given image and it is channel encoded before it is sent over a noisy channel to the decoder. At the decoder the semantic bit stream is channel decoded and use it as an input to autoencoder decoder to get the desired image at the decoder. A common knowledge of the images is shared between both the encoder and the decoder when the auto encoder is trained.

Scope and Information for Participants:

The objective of this tutorial is to produce an overview of the historical development and current status of semantic communication while providing an understanding on the challenges and possible future direction in the field. We expect to present our new findings on autoencoder based semantic communication for image transmission over error prone channels. It is expected to discuss the autoencoders in detail and their applications and how they can be used for semantic communications. Furthermore, it will be discussed how autoencoders can be trained for semantic communications and how they can be used in image transmission over a noisy channel. Challenges faced in transmitting the images and how machine to machine (M2M) image communications can be happened with the proposed architecture will also be discussed.

We expect to present our study providing a comprehensive overview on the historical development of AI algorithms in stock price prediction. In addition, through this tutorial it is expected to produce an in-depth overview on the kinds of algorithms that are currently utilized by researchers to generate next day price and price direction prediction based on numeric and textual data.

Highlights:

The conventional approach to communication involves transmitting the minimum number of bits with minimal errors between two points. This was based on Shannon's 1948 paper, which established channel capacity and showed that rates below capacity are possible without incurring an exponentially higher number of errors at the receiver side. However, this approach does not explicitly exploit the information about the source available at the transmitter side. Semantic communication is a paradigm that aims to address the second layer of communication, the semantic problem, by delivering the semantic meaning of the message instead of the exact form of the message. Currently, there is no unique transmission strategy for semantic communication, so a system must be designed in accordance with the current communication framework. The challenge is to ensure that the transmitter's semantic information is preserved at the receiver while transmitting through the physical channel, and research is required for different media types over conventional communication standards with semantic communications. This research aims to develop an autoencoder based semantic communication system to transmit images over a channel to optimize the bandwidth while maintaining the quality of the image. A multi layered autoencoder is used to generate a latent vector for a given image and it is channel encoded before it is sent over a noisy channel to the decoder. At the decoder the semantic bit stream is channel decoded and use it as an input to autoencoder decoder to get the desired image at the decoder. A common knowledge of the image is shared between both the encoder and the decoder when the auto encoder is trained. The objective of this workshop is to produce an overview of the historical development and current status of semantic communication while providing an understanding on the challenges and possible future direction in the field.

Access to Workshop:

CONF-MLA 2023 Workshop Glasgow - YouTube

Venue:

LT304, Department of Computer and Information Sciences, University of Strathclyde, Glasgow G1 1XQ, UK

VISA:

Welcome to GOV.UK

In order to ensure the information is correct and up to date, there may be changes which we are not aware of. And different countries have different rules for the visa application. It is always a good idea to check the latest regulations in your country. This page just gives some general information of the visa application.

UK Visa Information

What you need to do

• Check if what you plan to do in the UK is allowed as a Standard Visitor.
• Check you meet the eligibility requirements.
• Check if you need to apply for a visa to visit the UK.
• Apply for a Standard Visitor visa online - if you need one.

Check you meet the eligibility requirements

You must have a passport or travel document to enter the UK. It should be valid for the whole of your stay.

You must be able to show that:

• you'll leave the UK at the end of your visit
• you're able to support yourself and your dependants during your trip (or have funding from someone else to support you)
• you're able to pay for your return or onward journey (or have funding from someone else to pay for the journey)
• you'll not live in the UK for extended periods through frequent or successive visits, or make the UK your main home

Check if you need a visa to visit the UK

Depending on your nationality, you'll either:

• have to apply for a Standard Visitor visa before you travel to the UK
• be able to visit the UK for up to 6 months without needing a visa

You can check if you need a visa before you apply.
If you do not need a visa, you must still meet the Standard Visitor eligibility requirements to visit the UK. You may be asked questions at the UK border about your eligibility and the activities you plan to do.