Suchergebnisse

The precise mechanisms by which the information ecosystem polarizes society remain elusive. Focusing on political sorting in networks, we develop a computational model that examines how social network structure changes when individuals participate in information cascades, evaluate their behavior, and potentially rewire their connections to others as a result. Individuals follow proattitudinal information sources but are more likely to first hear and react to news shared by their social ties and only later evaluate these reactions by direct reference to the coverage of their preferred source. Reactions to news spread through the network via a complex contagion. Following a cascade, individuals who determine that their participation was driven by a subjectively "unimportant" story adjust their social ties to avoid being misled in the future. In our model, this dynamic leads social networks to politically sort when news outlets differentially report on the same topic, even when individuals do not know others' political identities. Observational follow network data collected on Twitter support this prediction: We find that individuals in more polarized information ecosystems lose cross-ideology social ties at a rate that is higher than predicted by chance. Importantly, our model reveals that these emergent polarized networks are less efficient at diffusing information: Individuals avoid what they believe to be "unimportant" news at the expense of missing out on subjectively "important" news far more frequently. This suggests that "echo chambers"—to the extent that they exist—may not echo so much as silence.

In this talk I introduce Transcendental Information Cascades, a method to understand the temporal dynamics of naturally occurring complex systems through the lens of a specific kind of spatio-temporal network that represents information token recurrence. The method suits very well for exploratory data analysis when it is unknown what is going on at the very low levels of a system's dynamics, and allows switching between looking at individual events or accumulated event sequences. I will present one application of this method in a digital humanities project where we analysed the entire corpus of Charles Dickens's novels and discuss the potential it has for other application areas such as the analysis of brain wave recordings, patent mining or the study of online communities for example. These are the slides of a talk at the RWTH Aachen University, Chair for Computational Social Sciences and Humanities (Germany, date of the talk 06-03-2018) {"references": ["Popper, K., 2013. Knowledge and the Body-Mind Problem: In defence of interaction. Routledge.", "Berners-Lee, Tim; Mark Fischetti (1999). Weaving the Web: The Original Design and Ultimate Destiny of the World Wide Web by its inventor. Britain: Orion Business. ISBN 0-7528-2090-7.", "Luczak-Roesch, M., Tinati, R., Simperl, E., Van Kleek, M., Shadbolt, N., & Simpson, R. (2014). Why won't aliens talk to us? Content and community dynamics in online citizen science. Proceedings of the Eighth AAAI Conference on Weblogs and Social Media, {ICWSM} 2014, Ann Arbor, Michigan, USA, June 1-4, 2014.", "Bikhchandani, Sushil, David Hirshleifer, and Ivo Welch. \"A theory of fads, fashion, custom, and cultural change as informational cascades.\" Journal of political Economy (1992): 992-1026.", "Cheng, Justin, et al. \"Can cascades be predicted?.\" Proceedings of the 23rd international conference on World wide web. International World Wide Web Conferences Steering Committee, 2014.", "Markus Luczak-Roesch, Ramine Tinati, and Nigel Shadbolt. 2015. When Resources Collide: Towards ...

Abstract Through a content analysis of 8,800 news items and six months of front pages of three Brazilian newspapers, all dealing with corruption and political transgression, the present article documents the remarkable bias of media coverage toward corruption scandals. Said bias is examined as an informational phenomenon, arising from key systemic and commercial factors of Brazil's news media: an information cascade of news on corruption formed, destabilizing the governing coalition and legitimizing the impeachment process of Dilma Rousseff. As this process gained momentum, questions of accountability were disregarded by the media, with harmful effects for democracy.

International audience ; This paper deals with a survey corpus. We present information retrieval about the speaker. We used finite state transducer cascades and we present here detailed results with an evaluation. This work is part of a French project to enhance the corpus ESLO (sociolinguistic survey taken in the city of Orléans). This survey has been realized in 1968 and the project is to save records in computer format, to transcribe them and to increase the transcription with annotations in XML format. This work was supported by a French ANR contract (ANR-06-CORP-023) and by European fund from Région Centre (FEDER). The corpus represent a collection of 200 interviews with the questions about the life in the city of Orléans: How long have you lived in Orléans for?, What led you to live in Orléans?, Do you like living in Orléans?, etc. and questions about the occupation or the family of the speaker, completed by recordings within a professional or private context. The recording situations are different: interviews, discussions between friends, recordings in microphone hidden, interviews with the political, academic and religious personalities, conversations between a social worker and parents in Psycho Medical Center of Orleans. In total, we have 300 hours of speech estimated to 4,500,000 words. More precisely, we worked on almost 120 transcribed hours representing 112 Transcriber XML files and 32 577 Kb. We worked on 105 files (31 004 Kb) and we evaluated the results on 7 files (1 573 Kb-5.1%). The transcription files have no punctuation marks, but the first letter of proper names is capitalized and acronyms are fully capitalized. We used the CasSys system (Friburger, Maurel, 2004) that computes texts with transducer cascades (Abney, 1996). The cascades we used are hand built: each transducer describes a local grammar for the recognition of some entities. Some times this recognition needs the succession of two or more transducers, in a specific order. More precisely, we used two cascades; the first one, for ...

International audience ; This paper deals with a survey corpus. We present information retrieval about the speaker. We used finite state transducer cascades and we present here detailed results with an evaluation. This work is part of a French project to enhance the corpus ESLO (sociolinguistic survey taken in the city of Orléans). This survey has been realized in 1968 and the project is to save records in computer format, to transcribe them and to increase the transcription with annotations in XML format. This work was supported by a French ANR contract (ANR-06-CORP-023) and by European fund from Région Centre (FEDER). The corpus represent a collection of 200 interviews with the questions about the life in the city of Orléans: How long have you lived in Orléans for?, What led you to live in Orléans?, Do you like living in Orléans?, etc. and questions about the occupation or the family of the speaker, completed by recordings within a professional or private context. The recording situations are different: interviews, discussions between friends, recordings in microphone hidden, interviews with the political, academic and religious personalities, conversations between a social worker and parents in Psycho Medical Center of Orleans. In total, we have 300 hours of speech estimated to 4,500,000 words. More precisely, we worked on almost 120 transcribed hours representing 112 Transcriber XML files and 32 577 Kb. We worked on 105 files (31 004 Kb) and we evaluated the results on 7 files (1 573 Kb-5.1%). The transcription files have no punctuation marks, but the first letter of proper names is capitalized and acronyms are fully capitalized. We used the CasSys system (Friburger, Maurel, 2004) that computes texts with transducer cascades (Abney, 1996). The cascades we used are hand built: each transducer describes a local grammar for the recognition of some entities. Some times this recognition needs the succession of two or more transducers, in a specific order. More precisely, we used two cascades; the first one, for ...

International audience ; This paper deals with a survey corpus. We present information retrieval about the speaker. We used finite state transducer cascades and we present here detailed results with an evaluation. This work is part of a French project to enhance the corpus ESLO (sociolinguistic survey taken in the city of Orléans). This survey has been realized in 1968 and the project is to save records in computer format, to transcribe them and to increase the transcription with annotations in XML format. This work was supported by a French ANR contract (ANR-06-CORP-023) and by European fund from Région Centre (FEDER). The corpus represent a collection of 200 interviews with the questions about the life in the city of Orléans: How long have you lived in Orléans for?, What led you to live in Orléans?, Do you like living in Orléans?, etc. and questions about the occupation or the family of the speaker, completed by recordings within a professional or private context. The recording situations are different: interviews, discussions between friends, recordings in microphone hidden, interviews with the political, academic and religious personalities, conversations between a social worker and parents in Psycho Medical Center of Orleans. In total, we have 300 hours of speech estimated to 4,500,000 words. More precisely, we worked on almost 120 transcribed hours representing 112 Transcriber XML files and 32 577 Kb. We worked on 105 files (31 004 Kb) and we evaluated the results on 7 files (1 573 Kb-5.1%). The transcription files have no punctuation marks, but the first letter of proper names is capitalized and acronyms are fully capitalized. We used the CasSys system (Friburger, Maurel, 2004) that computes texts with transducer cascades (Abney, 1996). The cascades we used are hand built: each transducer describes a local grammar for the recognition of some entities. Some times this recognition needs the succession of two or more transducers, in a specific order. More precisely, we used two cascades; the first one, for ...

International audience ; This paper deals with a survey corpus. We present information retrieval about the speaker. We used finite state transducer cascades and we present here detailed results with an evaluation. This work is part of a French project to enhance the corpus ESLO (sociolinguistic survey taken in the city of Orléans). This survey has been realized in 1968 and the project is to save records in computer format, to transcribe them and to increase the transcription with annotations in XML format. This work was supported by a French ANR contract (ANR-06-CORP-023) and by European fund from Région Centre (FEDER). The corpus represent a collection of 200 interviews with the questions about the life in the city of Orléans: How long have you lived in Orléans for?, What led you to live in Orléans?, Do you like living in Orléans?, etc. and questions about the occupation or the family of the speaker, completed by recordings within a professional or private context. The recording situations are different: interviews, discussions between friends, recordings in microphone hidden, interviews with the political, academic and religious personalities, conversations between a social worker and parents in Psycho Medical Center of Orleans. In total, we have 300 hours of speech estimated to 4,500,000 words. More precisely, we worked on almost 120 transcribed hours representing 112 Transcriber XML files and 32 577 Kb. We worked on 105 files (31 004 Kb) and we evaluated the results on 7 files (1 573 Kb-5.1%). The transcription files have no punctuation marks, but the first letter of proper names is capitalized and acronyms are fully capitalized. We used the CasSys system (Friburger, Maurel, 2004) that computes texts with transducer cascades (Abney, 1996). The cascades we used are hand built: each transducer describes a local grammar for the recognition of some entities. Some times this recognition needs the succession of two or more transducers, in a specific order. More precisely, we used two cascades; the first one, for ...

We analyze information diffusion using empirical data that tracks online communication around two instances of mass political mobilization that took place in Spain in 2011 and 2012. We also analyze protest-related communications during the year that elapsed between those protests. We compare the global properties of the topological and dynamic networks through which communication took place, as well as local changes in network composition. We show that changes in network structure underlie aggregated differences on how information diffused: an increase in network hierarchy is accompanied by a reduction in the average size of cascades. The increasing hierarchy affects not only the underlying communication topology but also the more dynamic structure of information exchange ; the increase is especially noticeable amongst certain categories of nodes (or users). Our findings suggest that the relationship between the structure of networks and their function in diffusing information is not as straightforward as some theoretical models of diffusion in networks imply.

This paper responds to the Alpine Rendez-Vous (ARV) 'crisis' in technology enhanced learning (TEL). It takes a contested area of policy, rapid change in the National Health Service (NHS), and documents the responses to 'information overload' by group of General Practitioners Practices in the North of England. Located between the spaces identified by Traxler and Lally as 'competitive industrialisation' and Web 1.0, and the consumer/ customer focus and ubiquitous ownership enabled by portable and devices and web 2.0, in this work we see the parallels of the responses of publicly funded bodies moving towards privatisation as part of a neo-liberal agenda. Interviews with health professionals revealed marginalized spaces for informal learning in their workplaces; and a desire to build a community that would enable them to overcome the time/space barriers to networking. The EU Learning Layers Integrating Project develops mobile and social technologies that unlock and enable peer production within and across traditional workplace boundaries. Through the health professional narratives, we capture insights into their daily life, enable the articulation of their needs for an online 'Help-Seeking' networking service, underpinned by their desire to consult what Vygotsky calls 'the more capable peer'.

The pandemic of the 2019 novel Coronavirus has seen unprecedented exponential growth. Within three months, 192 countries have been affected, crossing more than 1 million confirmed cases and over 60 thousand deaths until the first week of April. Decision making in such a pandemic becomes difficult due to limited data on the nature of the disease and its propagation, course, prevention, and treatment. The pandemic response has varied from country to country and has resulted in a heterogeneous timeline for novel Coronavirus propagation. We compared the public health measures taken by various countries and the potential impact on the spread. We studied 6 countries including China, Italy, South Korea, Singapore, United Kingdom(UK), United States(US), and the special administrative region of Hong Kong. All articles, press releases, and websites of government entities published over a five-month period were included. A comparison of the date of the first diagnosed case, the spread of disease, and time since the first case and major public health policy implemented for prevention and containment and current cases was done. An emphasis on early and aggressive border restriction and surveillance of travelers from infected areas, use of information technology, and social distancing is necessary for control of the novel pandemic. Moving forwards, improvement in infrastructure, and adequate preparedness for pandemics is required.

Modeling online discourse dynamics is a core activity in understanding the spread of information, both offline and online, and emergent online behavior. There is currently a disconnect between the practitioners of online social media analysis -- usually social, political and communication scientists -- and the accessibility to tools capable of examining online discussions of users. Here we present evently, a tool for modeling online reshare cascades, and particularly retweet cascades, using self-exciting processes. It provides a comprehensive set of functionalities for processing raw data from Twitter public APIs, modeling the temporal dynamics of processed retweet cascades and characterizing online users with a wide range of diffusion measures. This tool is designed for researchers with a wide range of computer expertise, and it includes tutorials and detailed documentation. We illustrate the usage of evently with an end-to-end analysis of online user behavior on a topical dataset relating to COVID-19. We show that, by characterizing users solely based on how their content spreads online, we can disentangle influential users and online bots.

ABSTRACT:Future projections of coastal erosion, which are one of the most demanded climate services in coastal areas, are mainly developed using top-down approaches. These approaches consist of undertaking a sequence of steps that include selecting emission or concentration scenarios and climate models, correcting models bias, applying downscaling methods, and implementing coastal erosion models. The information involved in this modelling chain cascades across steps, and so does related uncertainty, which accumulates in the results. Here, we develop long-term multi-ensemble probabilistic coastal erosion projections following the steps of the top-down approach, factorise, decompose and visualise the uncertainty cascade using real data and analyse the contribution of the uncertainty sources (knowledge-based and intrinsic) to the total uncertainty. We find a multi-modal response in long-term erosion estimates and demonstrate that not sampling internal climate variability?s uncertainty sufficiently could lead to a truncated outcomes range, affecting decision-making. Additionally, the noise arising from internal variability (rare outcomes) appears to be an important part of the full range of results, as it turns out that the most extreme shoreline retreat events occur for the simulated chronologies of climate forcing conditions. We conclude that, to capture the full uncertainty, all sources need to be properly sampled considering the climate-related forcing variables involved, the degree of anthropogenic impact and time horizon targeted. ; AT acknowledges the financial support from the FENIX Project by the Government of Cantabria. This research was also funded by the Spanish Government via the grant RISKCOADAPT (BIA2017-89401-R).

Automated detection of the content of images remains a challenging problem in artificial intelligence. Hence, continuous manual monitoring of restricted development zones is critical to maintaining territorial integrity and national security. In this regard, local governments of the Republic of Korea conduct four periodic inspections per year to preserve national territories from illegal encroachments and unauthorized developments in restricted zones. The considerable expense makes responding to illegal developments difficult for local governments. To address this challenge, we propose a deep-learning-based Cascade Mask region-based convolutional neural network (R-CNN) algorithm designed to perform automated detection of greenhouses in aerial photographs for efficient and continuous monitoring of restricted development zones in the Republic of Korea. Our proposed model is regional-based because it was optimized for the Republic of Korea via transfer learning and hyperparameter tuning, which improved the efficiency of the automated detection of greenhouse facilities. The experimental results demonstrated that the mAP value of the proposed Cascade Mask R-CNN model was 83.6, which was 12.83 higher than baseline mask R-CNN, and 0.9 higher than Mask R-CNN with hyperparameter tuning and transfer learning considered. Similarly, the F1-score of the proposed Cascade Mask R-CNN model was 62.07, which outperformed those of the baseline mask R-CNN and the Mask R-CNN with hyperparameter tuning and transfer learning considered (i.e., the F1-score 52.33 and 59.13, respectively). The proposed improved Cascade Mask R-CNN model is expected to facilitate efficient and continuous monitoring of restricted development zones through routine screening procedures. Moreover, this work provides a baseline for developing an integrated management system for national-scale land-use planning and development infrastructure by synergizing geographical information systems, remote sensing, and deep learning models.

Automated detection of the content of images remains a challenging problem in artificial intelligence. Hence, continuous manual monitoring of restricted development zones is critical to maintaining territorial integrity and national security. In this regard, local governments of the Republic of Korea conduct four periodic inspections per year to preserve national territories from illegal encroachments and unauthorized developments in restricted zones. The considerable expense makes responding to illegal developments difficult for local governments. To address this challenge, we propose a deep-learning-based Cascade Mask region-based convolutional neural network (R-CNN) algorithm designed to perform automated detection of greenhouses in aerial photographs for efficient and continuous monitoring of restricted development zones in the Republic of Korea. Our proposed model is regional-based because it was optimized for the Republic of Korea via transfer learning and hyperparameter tuning, which improved the efficiency of the automated detection of greenhouse facilities. The experimental results demonstrated that the mAP value of the proposed Cascade Mask R-CNN model was 83.6, which was 12.83 higher than baseline mask R-CNN, and 0.9 higher than Mask R-CNN with hyperparameter tuning and transfer learning considered. Similarly, the F1-score of the proposed Cascade Mask R-CNN model was 62.07, which outperformed those of the baseline mask R-CNN and the Mask R-CNN with hyperparameter tuning and transfer learning considered (i.e., the F1-score 52.33 and 59.13, respectively). The proposed improved Cascade Mask R-CNN model is expected to facilitate efficient and continuous monitoring of restricted development zones through routine screening procedures. Moreover, this work provides a baseline for developing an integrated management system for national-scale land-use planning and development infrastructure by synergizing geographical information systems, remote sensing, and deep learning models.

The demand for the sensor-based detection of camouflage objects widely exists in biological research, remote sensing, and military applications. However, the performance of traditional object detection algorithms is limited, as they are incapable of extracting informative parts from low signal-to-noise ratio features. To address this problem, we propose Camouflaged Object Detection with Cascade and Feedback Fusion (CODCEF), a deep learning framework based on an RGB optical sensor that leverages a cascaded structure with Feedback Partial Decoders (FPD) instead of a traditional encoder–decoder structure. Through a selective fusion strategy and feedback loop, FPD reduces the loss of information and the interference of noises in the process of feature interweaving. Furthermore, we introduce Pixel Perception Fusion (PPF) loss, which aims to pay more attention to local pixels that might become the edges of an object. Experimental results on an edge device show that CODCEF achieved competitive results compared with 10 state-of-the-art methods.