Suchergebnisse

This short paper presents what is currently the main challenge associated with using Generative Adversarial Networks (GANs) to generate visual media. That challenge is around automatically determining the authenticity of an image/video, which is of increasing importance as fake videos and images start to proliferate on social media, especially when associated with political campaigning. The paper introduces GANs, outlines how they are used to generate visual media and summarises this major challenge.

Open Access is a movement within scholarly research across all disciplines whereby the outputs, including publications and data, are made freely available to everyone without restriction. In the case of Open Access to data, it also covers data from public bodies like local governments. In Ireland we are slowly making progress with open access to publications and poor progress to research data while our Open Access to data from public bodies is progressing well

Ireland's greatest natural resource, apart from its people, is its ocean territory, comprising 220,000,000 acres and representing ten times our landmass. Our oceans are rich in natural resources and hold huge potential for energy generation. Yet our oceans face threats, some of which can be monitored, and perhaps addressed, using ICT. In this presentation I give an overview of one small part of our ocean, Galway Bay, and I introduce the SmartBay project, a pilot which brings together industry (both large and small companies, indigenous and multi-national) researchers, and government agencies into a platform project for developing ocean technologies. I describe the work on one kind of sensor developed within our research centre CLARITY, and describe how it can be used to monitor phosphate emissions from lakes and rivers as they ingress into Galway Bay and how this in turn can lead to the development of harmful algal blooms within the bay area.

The body of content available on Twitter undoubtedly contains a diverse range of political insight and commentary. But, to what extent is this representative of an electorate? Can we model political sentiment effectively enough to capture the voting intentions of a nation during an election capaign? We use the recent Irish General Election as a case study for investigating the potential to model political sentiment through mining of social media. Our approach combines sentiment analysis using supervised learning and volume-based measures. We evaluate against the conventional election polls and the final election result. We find that social analytics using both volume-based measures and sentiment analysis are predictive and wemake a number of observations related to the task of monitoring public sentiment during an election campaign, including examining a variety of sample sizes, time periods as well as methods for qualitatively exploring the underlying content.

There are many applications of Generative Adversarial Networks (GANs) in fields like computer vision, natural language processing, speech synthesis, and more. Undoubtedly the most notable results have been in the area of image synthesis and in particular in the generation of deepfake videos. While deepfakes have received much negative media coverage, they can be a useful technology in applications like entertainment, customer relations, or even assistive care. One problem with generating deepfakes is the requirement for a lot of image training data of the subject which is not an issue if the subject is a celebrity for whom many images already exist. If there are only a small number of training images then the quality of the deepfake will be poor. Some media reports have indicated that a good deepfake can be produced with as few as 500 images but in practice, quality deepfakes require many thousands of images, one of the reasons why deepfakes of celebrities and politicians have become so popular. In this study, we exploit the property of a GAN to produce images of an individual with variable facial expressions which we then use to generate a deepfake. We observe that with such variability in facial expressions of synthetic GAN-generated training images and a reduced quantity of them, we can produce a near-realistic deepfake videos.

Water management is an important part of monitoring the natural environment and includes monitoring water quality of both coastal and inland marine locations. This covers the detection of pollution and monitoring the development of harmful algal blooms as well as coastal features and wave patterns. For many years water managers relied on field measurements for coastal monitoring and water quality evaluation. This type of sampling is quite limited on both temporal and spatial scales and is ineffective for capturing dynamic marine events, essential for increased knowledge and better decision making. It also involves costly, time and labour-intensive on-site sampling and data collection. The introduction of new policies such as the EU Water Framework Directive has increased pressure on governments to adopt new methods for continuous monitoring of all water bodies. In recent years, the use of in-situ wireless sensor networks (WSNs) for marine environmental monitoring has been developing to allow continuous real-time monitoring of the marine environment at greater temporal and spatial resolutions. WSNs have brought about great advantages and increased opportunities for environmental monitoring. The WSN concept envisages a world of ubiquitous sensing through large scale deployments of self-sustaining WSNs linked to digital communications, continuously monitoring our environment and detecting and reporting changes in its quality. However there are still several challenges remaining in the area of in-situ wireless sensor networks in order to realise this vision. Among these is the issue of data reliability and the need for a sensor network that adapts to changes in the availability and reliability of its own sensors, as it monitors an already changing environment. We are developing a reputation and trust based event detection system for environmental monitoring which incorporates alternative sensing modalities such as visual sensors (e.g. digital cameras and satellite sensing) and context information alongside an ...

The decisions of the European Parliament (EP) are shown to influence both EU emission allowance (EUA) prices and volatility. Reductions in price and increases in volatility are observed when EP decisions are (i) not "party-political" in origin, (ii) made during times of low market sentiment, or (iii) made during times of low market attention. Daily EUA prices from 2007 to 2014 are used in the study, with decisions analysed using an event study approach for price impact, and a GARCH specification for volatility impact. Our findings suggest the need for policymakers to improve communication of long-term strategies for the EUA market. This aims to reduce the evident ongoing uncertainty experienced by traders around each decision made by the EP. The finding that sentiment and market attention at the time of an EP decision influences the market's reaction indicates a need to consider market dynamics in terms of decision timing, so that market turbulence is not an unintended by-product of an EP decision. Indeed some form of medium term forward guidance may be called for.

Water management is an important part of monitoring the natural environment and includes monitoring water quality of both coastal and inland marine locations. This covers the detection of pollution and monitoring the development of harmful algal blooms as well as coastal features and wave patterns. For many years water managers relied on field measurements for coastal monitoring and water quality evaluation. This type of sampling is quite limited on both temporal and spatial scales and is ineffective for capturing dynamic marine events, essential for increased knowledge and better decision making. It also involves costly, time and labour-intensive on-site sampling and data collection. The introduction of new policies such as the EU Water Framework Directive has increased pressure on governments to adopt new methods for continuous monitoring of all water bodies. In recent years, the use of in-situ wireless sensor networks (WSNs) for marine environmental monitoring has been developing to allow continuous real-time monitoring of the marine environment at greater temporal and spatial resolutions. WSNs have brought about great advantages and increased opportunities for environmental monitoring. The WSN concept envisages a world of ubiquitous sensing through large scale deployments of self-sustaining WSNs linked to digital communications, continuously monitoring our environment and detecting and reporting changes in its quality. However there are still several challenges remaining in the area of in-situ wireless sensor networks in order to realise this vision. Among these is the issue of data reliability and the need for a sensor network that adapts to changes in the availability and reliability of its own sensors, as it monitors an already changing environment. We are developing a reputation and trust based event detection system for environmental monitoring which incorporates alternative sensing modalities such as visual sensors (e.g. digital cameras and satellite sensing) and context information alongside an ...

Water management is an important part of monitoring the natural environment and includes monitoring water quality of both coastal and inland marine locations. This covers the detection of pollution and monitoring the development of harmful algal blooms as well as coastal features and wave patterns. For many years water managers relied on field measurements for coastal monitoring and water quality evaluation. This type of sampling is quite limited on both temporal and spatial scales and is ineffective for capturing dynamic marine events, essential for increased knowledge and better decision making. It also involves costly, time and labour-intensive on-site sampling and data collection. The introduction of new policies such as the EU Water Framework Directive has increased pressure on governments to adopt new methods for continuous monitoring of all water bodies. In recent years, the use of in-situ wireless sensor networks (WSNs) for marine environmental monitoring has been developing to allow continuous real-time monitoring of the marine environment at greater temporal and spatial resolutions. WSNs have brought about great advantages and increased opportunities for environmental monitoring. The WSN concept envisages a world of ubiquitous sensing through large scale deployments of self-sustaining WSNs linked to digital communications, continuously monitoring our environment and detecting and reporting changes in its quality. However there are still several challenges remaining in the area of in-situ wireless sensor networks in order to realise this vision. Among these is the issue of data reliability and the need for a sensor network that adapts to changes in the availability and reliability of its own sensors, as it monitors an already changing environment. We are developing a reputation and trust based event detection system for environmental monitoring which incorporates alternative sensing modalities such as visual sensors (e.g. digital cameras and satellite sensing) and context information alongside an ...

Polls have been used for decades as a tool to gauge public opinion on specified topics. Originally used in US Presidential elections, they are now used to gather point-of-time information on politicians, political issues, brand names, products, even prospective storylines in movies. Robust polls typically specify a priori an acceptable level of sampling error e.g. +/- 3% in order to calculate the size of the random sample needed. While these error rates are acceptable for most cases, having to manually poll a random sample of up to 1,000 people means opinion polls are expensive to carry out. In order to identify secular trends, polling must be repeated at frequent intervals, which makes monitoring for trends even more expensive. Surveys can prove more cost-effective, but still require resources to recruit participants and collate results, with the risk of latency from time of issuing a survey to aggregation of results. For this reason we look to online social media as a potential low-cost, continuous and scalable alternative source for opinion mining that can be readily replicated.

In a lifelog, data from various sources are combined to form a record from which one can retrieve information about oneself and the environment in which one is situated. It could be considered similar to an automated biography. Lifelog technology is still at an early stage of development. However, the history of lifelogs so far shows a clear academic, corporate and governmental interest. Therefore, a thorough inquiry into the ethical aspects of lifelogs could prove beneficial to the responsible development of this field. This article maps the main ethically relevant challenges and opportunities associated with the further development of lifelog technologies as discussed in the scholarly literature. By identifying challenges and opportunities in the current debate, we were able to identify other challenges and opportunities left unmentioned. Some of these challenges are partly explained by a blind spot in the current debate. Whilst the current debate focuses mainly on lifelogs held by individuals, lifelogs held by governmental institutions and corporations pose idiosyncratic ethical concerns as well. We have provided a brief taxonomy of lifelog technology to show the variety in uses for lifelogs. In addition, we provided a general approach to alleviate the ethical challenges identified in the critical analysis.

Demands from various scientific and management communities along with legislative requirements at national and international levels have led to a need for innovative research into large-scale, low-cost, reliable monitoring of our marine and freshwater environments. In this paper we demonstrate the benefits of a multi-modal approach to monitoring and how an in-situ sensor network can be enhanced with the use of contextual image data. We provide an outline of the deployment of a visual sensing system at a busy port and the need for monitoring shipping traffic at the port. Subsequently we present an approach for detecting ships in a challenging image data set and discuss how this can help to create an intelligent marine monitoring network.

In this work we discuss a new, but highly relevant, topic to the multimedia community; systems to inform individuals of their carbon footprint, which could ultimately effect change in community carbon footprint-related activities. The reduction of carbon emissions is now an important policy driver of many governments, and one of the major areas of focus is in reducing the energy demand from the consumers i.e. all of us individually. In terms of CO2 generated from energy consumption, there are three predominant factors, namely electricity usage, thermal related costs, and transport usage. Standard home electricity and heating sensors can be used to measure the former two aspects, and in this paper we evaluate a novel technique to estimate an individual's transport-related carbon emissions through the use of a simple wearable accelerometer. We investigate how providing this novel estimation of transport-related carbon emissions through an interactive web site and mobile phone app engages a set of users in becoming more aware of their carbon emissions. Our evaluations involve a group of 6 users collecting 25 million accelerometer readings and 12.5 million power readings vs. a control group of 16 users collecting 29.7 million power readings.

International audience ; The TREC Video Retrieval Evaluation (TRECVID) 2014 was a TREC-style video analysis and retrieval evaluation, the goal of which remains to promote progress in content-based exploitation of digital video via open, metrics-based evaluation. Over the last dozen years this effort has yielded a better under- standing of how systems can effectively accomplish such processing and how one can reliably benchmark their performance. TRECVID is funded by the NIST with support from other US government agencies. Many organizations and individuals worldwide contribute significant time and effort.