Search

BACKGROUND: Predictive biomarkers of immune checkpoint inhibitor (ICI) efficacy are currently lacking for non-small cell lung cancer (NSCLC). Here, we describe the results from the Anti-PD-1 Response Prediction DREAM Challenge, a crowdsourced initiative that enabled the assessment of predictive models by using data from two randomized controlled clinical trials (RCTs) of ICIs in first-line metastatic NSCLC.METHODS: Participants developed and trained models using public resources. These were evaluated with data from the CheckMate 026 trial (NCT02041533), according to the model-to-data paradigm to maintain patient confidentiality. The generalizability of the models with the best predictive performance was assessed using data from the CheckMate 227 trial (NCT02477826). Both trials were phase III RCTs with a chemotherapy control arm, which supported the differentiation between predictive and prognostic models. Isolated model containers were evaluated using a bespoke strategy that considered the challenges of handling transcriptome data from clinical trials.RESULTS: A total of 59 teams participated, with 417 models submitted. Multiple predictive models, as opposed to a prognostic model, were generated for predicting overall survival, progression-free survival, and progressive disease status with ICIs. Variables within the models submitted by participants included tumor mutational burden (TMB), programmed death ligand 1 (PD-L1) expression, and gene-expression-based signatures. The best-performing models showed improved predictive power over reference variables, including TMB or PD-L1.CONCLUSIONS: This DREAM Challenge is the first successful attempt to use protected phase III clinical data for a crowdsourced effort towards generating predictive models for ICI clinical outcomes and could serve as a blueprint for similar efforts in other tumor types and disease states, setting a benchmark for future studies aiming to identify biomarkers predictive of

Machine learning methodologies can provide viable solutions to a wide range of complex real-world problems. However, in practice, issues such as data noise, class imbalance, and uncertainty inherent in datasets and models are frequently encountered. This dissertation investigates machine learning strategies for crowdsourced labeling, medical diagnosis, driver distraction detection, and biological image classification. It proposes a soft-weighted aggregation algorithm for crowdsourced labels that uses annotators' consistency in relation to a reference classifier to determine their reliability, demonstrating a 20% improvement in accuracy over ten existing label aggregation techniques across ten distinct datasets when only three workers are present. The dissertation also describes a hierarchical classification methodology that uses disease taxonomy to improve chest radiograph classifications, demonstrating a 15% improvement in accuracy and AUC score across different diseases when the proposed approaches are used. It presents a cascaded multi-planar convolutional neural network design for improved thalamic nuclei segmentation in MRI scans. Transfer learning is used in two different applications: identifying driver distractions in images and classifying primary cilia in microscopy data. The efficacy of these methodologies is supported by experimental results on real-world datasets. The approaches and findings in this dissertation contribute to machine learning advancements and may provide practical solutions to some of the challenges encountered in various sectors, such as data science, radiology, and road safety.

The existence of a link between the gut microbiome and autism spectrum disorder (ASD) is well established in mice, but in human populations, efforts to identify microbial biomarkers have been limited due to a lack of appropriately matched controls, stratification of participants within the autism spectrum, and sample size. To overcome these limitations, we crowdsourced the recruitment of families with age-matched sibling pairs between 2 and 7 years old (within 2 years of each other), where one child had a diagnosis of ASD and the other did not. Parents collected stool samples, provided a home video of their ASD child's natural social behavior, and responded online to diet and behavioral questionnaires. 16S rRNA V4 amplicon sequencing of 117 samples (60 ASD and 57 controls) identified 21 amplicon sequence variants (ASVs) that differed significantly between the two cohorts: 11 were found to be enriched in neurotypical children (six ASVs belonging to the Lachnospiraceae family), while 10 were enriched in children with ASD (including Ruminococcaceae and Bacteroidaceae families). Summarizing the expected KEGG orthologs of each predicted genome, the taxonomic biomarkers associated with children with ASD can use amino acids as precursors for butyragenic pathways, potentially altering the availability of neurotransmitters like glutamate and gamma aminobutyric acid (GABA).IMPORTANCE Autism spectrum disorder (ASD), which now affects 1 in 54 children in the United States, is known to have comorbidity with gut disorders of a variety of types; however, the link to the microbiome remains poorly characterized. Recent work has provided compelling evidence to link the gut microbiome to the autism phenotype in mouse models, but identification of specific taxa associated with autism has suffered replicability issues in humans. This has been due in part to sample size that sufficiently covers the spectrum of phenotypes known to autism (which range from subtle to severe) and a lack of a

The existence of a link between the gut microbiome and autism spectrum disorder (ASD) is well established in mice, but in human populations, efforts to identify microbial biomarkers have been limited due to a lack of appropriately matched controls, stratification of participants within the autism spectrum, and sample size. To overcome these limitations, we crowdsourced the recruitment of families with age-matched sibling pairs between 2 and 7 years old (within 2 years of each other), where one child had a diagnosis of ASD and the other did not. Parents collected stool samples, provided a home video of their ASD child's natural social behavior, and responded online to diet and behavioral questionnaires. 16S rRNA V4 amplicon sequencing of 117 samples (60 ASD and 57 controls) identified 21 amplicon sequence variants (ASVs) that differed significantly between the two cohorts: 11 were found to be enriched in neurotypical children (six ASVs belonging to the Lachnospiraceae family), while 10 were enriched in children with ASD (including Ruminococcaceae and Bacteroidaceae families). Summarizing the expected KEGG orthologs of each predicted genome, the taxonomic biomarkers associated with children with ASD can use amino acids as precursors for butyragenic pathways, potentially altering the availability of neurotransmitters like glutamate and gamma aminobutyric acid (GABA).IMPORTANCE Autism spectrum disorder (ASD), which now affects 1 in 54 children in the United States, is known to have comorbidity with gut disorders of a variety of types; however, the link to the microbiome remains poorly characterized. Recent work has provided compelling evidence to link the gut microbiome to the autism phenotype in mouse models, but identification of specific taxa associated with autism has suffered replicability issues in humans. This has been due in part to sample size that sufficiently covers the spectrum of phenotypes known to autism (which range from subtle to severe) and a lack of a

This data set was crowdsourced at the 2021 Helmholtz MT ARD ST3 meeting from attendants of the Machine Learning Tutorial on Sep 30, 2021. For more details on the event, see https://indico.desy.de/event/28823

Psoriasis is a skin condition that affects over 100 million people worldwide, while multiple myeloma (MM) accounts for 10% of all hematologic malignancies in the US. There has been limited research on the intersection of psoriasis and MM, and clinicians often face difficult decisions in treating patients diagnosed with both conditions. For instance, the management of psoriasis with systemic immunotherapies in MM patients can be challenging because of concern about immunosuppression and possible worsening of MM. Online crowdsourcing platforms have recently become innovative tools that can actively empower patients in scientific research by enabling the contribution of health data. One such platform, HealthTree®, helps MM patients find optimal myeloma treatments and has registered > 6000 patients, many of whom have uploaded medical records and genetic profiles. By taking advantage of patient health data available on HealthTree, researchers can gain a greater understanding of the clinical characteristics and treatment responses of patients diagnosed with psoriasis and MM. In this case series, we first report a psoriasis and MM patient treated with the IL-17 inhibitor ixekizumab who demonstrated a temporary, 2-month improvement in MM biomarkers (M-protein, kappa, and kappa:lambda ratio). We then report on the clinical characteristics of 18 patients with verified profiles on HealthTree indicating concurrent psoriasis and MM conditions. We surveyed gender, age, psoriasis type, psoriasis treatment history, myeloma type, myeloma genetic features, and myeloma association with bone damage, hypercalcemia, or osteopenia. Four patients were treated with systemic immunomodulators for psoriasis, with responses suggesting that these therapies did not worsen MM progression. Our results validate crowdsourcing as a way to assess patient demographics and treatment responses for use in dermatology research. We examine the demographics of patients diagnosed with psoriasis and MM and i

Scientists are increasingly using volunteer efforts of citizen scientists to classify images captured by motion-activated trail cameras. The rising popularity of citizen science reflects its potential to engage the public in conservation science and accelerate processing of the large volume of images generated by trail cameras. While image classification accuracy by citizen scientists can vary across species, the influence of other factors on accuracy is poorly understood. Inaccuracy diminishes the value of citizen science derived data and prompts the need for specific best-practice protocols to decrease error. We compare the accuracy between three programs that use crowdsourced citizen scientists to process images online: Snapshot Serengeti, Wildwatch Kenya, and AmazonCam Tambopata. We hypothesized that habitat type and camera settings would influence accuracy. To evaluate these factors, each photograph was circulated to multiple volunteers. All volunteer classifications were aggregated to a single best answer for each photograph using a plurality algorithm. Subsequently, a subset of these images underwent expert review and were compared to the citizen scientist results. Classification errors were categorized by the nature of the error (e.g., false species or false empty), and reason for the false classification (e.g., misidentification). Our results show that Snapshot Serengeti had the highest accuracy (97.9%), followed by AmazonCam Tambopata (93.5%), then Wildwatch Kenya (83.4%). Error type was influenced by habitat, with false empty images more prevalent in open-grassy habitat (27%) compared to woodlands (10%). For medium to large animal surveys across all habitat types, our results suggest that to significantly improve accuracy in crowdsourced projects, researchers should use a trail camera set up protocol with a burst of three consecutive photographs, a short field of view, and determine camera sensitivity settings based on in situ testing. Accuracy level comp

Psoriasis is a skin condition that affects over 100 million people worldwide, while multiple myeloma (MM) accounts for 10% of all hematologic malignancies in the US. There has been limited research on the intersection of psoriasis and MM, and clinicians often face difficult decisions in treating patients diagnosed with both conditions. For instance, the management of psoriasis with systemic immunotherapies in MM patients can be challenging because of concern about immunosuppression and possible worsening of MM. Online crowdsourcing platforms have recently become innovative tools that can actively empower patients in scientific research by enabling the contribution of health data. One such platform, HealthTree®, helps MM patients find optimal myeloma treatments and has registered > 6000 patients, many of whom have uploaded medical records and genetic profiles. By taking advantage of patient health data available on HealthTree, researchers can gain a greater understanding of the clinical characteristics and treatment responses of patients diagnosed with psoriasis and MM. In this case series, we first report a psoriasis and MM patient treated with the IL-17 inhibitor ixekizumab who demonstrated a temporary, 2-month improvement in MM biomarkers (M-protein, kappa, and kappa:lambda ratio). We then report on the clinical characteristics of 18 patients with verified profiles on HealthTree indicating concurrent psoriasis and MM conditions. We surveyed gender, age, psoriasis type, psoriasis treatment history, myeloma type, myeloma genetic features, and myeloma association with bone damage, hypercalcemia, or osteopenia. Four patients were treated with systemic immunomodulators for psoriasis, with responses suggesting that these therapies did not worsen MM progression. Our results validate crowdsourcing as a way to assess patient demographics and treatment responses for use in dermatology research. We examine the demographics of patients diagnosed with psoriasis and MM and i

Scientists are increasingly using volunteer efforts of citizen scientists to classify images captured by motion-activated trail cameras. The rising popularity of citizen science reflects its potential to engage the public in conservation science and accelerate processing of the large volume of images generated by trail cameras. While image classification accuracy by citizen scientists can vary across species, the influence of other factors on accuracy is poorly understood. Inaccuracy diminishes the value of citizen science derived data and prompts the need for specific best-practice protocols to decrease error. We compare the accuracy between three programs that use crowdsourced citizen scientists to process images online: Snapshot Serengeti, Wildwatch Kenya, and AmazonCam Tambopata. We hypothesized that habitat type and camera settings would influence accuracy. To evaluate these factors, each photograph was circulated to multiple volunteers. All volunteer classifications were aggregated to a single best answer for each photograph using a plurality algorithm. Subsequently, a subset of these images underwent expert review and were compared to the citizen scientist results. Classification errors were categorized by the nature of the error (e.g., false species or false empty), and reason for the false classification (e.g., misidentification). Our results show that Snapshot Serengeti had the highest accuracy (97.9%), followed by AmazonCam Tambopata (93.5%), then Wildwatch Kenya (83.4%). Error type was influenced by habitat, with false empty images more prevalent in open-grassy habitat (27%) compared to woodlands (10%). For medium to large animal surveys across all habitat types, our results suggest that to significantly improve accuracy in crowdsourced projects, researchers should use a trail camera set up protocol with a burst of three consecutive photographs, a short field of view, and determine camera sensitivity settings based on in situ testing. Accuracy level comp

Scientists are increasingly using volunteer efforts of citizen scientists to classify images captured by motion-activated trail cameras. The rising popularity of citizen science reflects its potential to engage the public in conservation science and accelerate processing of the large volume of images generated by trail cameras. While image classification accuracy by citizen scientists can vary across species, the influence of other factors on accuracy is poorly understood. Inaccuracy diminishes the value of citizen science derived data and prompts the need for specific best-practice protocols to decrease error. We compare the accuracy between three programs that use crowdsourced citizen scientists to process images online: Snapshot Serengeti, Wildwatch Kenya, and AmazonCam Tambopata. We hypothesized that habitat type and camera settings would influence accuracy. To evaluate these factors, each photograph was circulated to multiple volunteers. All volunteer classifications were aggregated to a single best answer for each photograph using a plurality algorithm. Subsequently, a subset of these images underwent expert review and were compared to the citizen scientist results. Classification errors were categorized by the nature of the error (e.g., false species or false empty), and reason for the false classification (e.g., misidentification). Our results show that Snapshot Serengeti had the highest accuracy (97.9%), followed by AmazonCam Tambopata (93.5%), then Wildwatch Kenya (83.4%). Error type was influenced by habitat, with false empty images more prevalent in open-grassy habitat (27%) compared to woodlands (10%). For medium to large animal surveys across all habitat types, our results suggest that to significantly improve accuracy in crowdsourced projects, researchers should use a trail camera set up protocol with a burst of three consecutive photographs, a short field of view, and determine camera sensitivity settings based on in situ testing. Accuracy level comp

[EN] The use of artificially intelligent techniques to overcome specific shortcomings within e-learning systems is a well-researched area that keeps on evolving in an attempt to optimise such resourceful practices. The lack of personalization and the sentiment of isolation coupled with a feeling of being treated like all others, tends to discourage and push learners away from courses that are very well prepared academically and excellently projected intellectually. The use of recommender systems to deliver relevant information in a timely manner that is specifically differentiated to a unique learner is once more being investigated to alievate the e-learning issue of being impersonal. The application of such a technique also assists the learner by reducing information overload and providing learning material that can be shared, criticized and reviewed at one’s own pace. In this paper we propose the use of a fully automated recommender system based on recent AI developments together with Web 2.0 applications and socially networked technologies. We argue that such technologies have provided the extra capabilities that were required to deliver a realistic and practical interfacing medium to assist online learners and take recommender systems to the next level.

The purpose of this thesis is to research how consumer recommendations can be used in terms of marketing and sales purposes in the Fast Moving Consumer Goods sector. Due to technological changes and a shift from traditional to digital marketing, the way of communicating to the audience has changed. Consumer recommendations, micro influencers and social media platforms are a part of digital marketing that is growing progressively. The paradigm of the thesis is interpretivism with following a qualitative research method, inductive reasoning and a case study in focus. The findings show that digital marketing, thanks to aspects as social media and e-commerce, has been growing, however has not completely taken over the marketing field. Influencers are getting a bigger role as marketing tools with the shift of power evolving. Due to these reasons is why and how consumer recommendations have become such a big trend nowadays.

[EN] The use of artificially intelligent techniques to overcome specific shortcomings within e-learning systems is a well-researched area that keeps on evolving in an attempt to optimise such resourceful practices. The lack of personalization and the sentiment of isolation coupled with a feeling of being treated like all others, tends to discourage and push learners away from courses that are very well prepared academically and excellently projected intellectually. The use of recommender systems to deliver relevant information in a timely manner that is specifically differentiated to a unique learner is once more being investigated to alievate the e-learning issue of being impersonal. The application of such a technique also assists the learner by reducing information overload and providing learning material that can be shared, criticized and reviewed at one’s own pace. In this paper we propose the use of a fully automated recommender system based on recent AI developments together with Web 2.0 applications and socially networked technologies. We argue that such technologies have provided the extra capabilities that were required to deliver a realistic and practical interfacing medium to assist online learners and take recommender systems to the next level.

[EN] The use of artificially intelligent techniques to overcome specific shortcomings within e-learning systems is a well-researched area that keeps on evolving in an attempt to optimise such resourceful practices. The lack of personalization and the sentiment of isolation coupled with a feeling of being treated like all others, tends to discourage and push learners away from courses that are very well prepared academically and excellently projected intellectually. The use of recommender systems to deliver relevant information in a timely manner that is specifically differentiated to a unique learner is once more being investigated to alievate the e-learning issue of being impersonal. The application of such a technique also assists the learner by reducing information overload and providing learning material that can be shared, criticized and reviewed at one’s own pace. In this paper we propose the use of a fully automated recommender system based on recent AI developments together with Web 2.0 applications and socially networked technologies. We argue that such technologies have provided the extra capabilities that were required to deliver a realistic and practical interfacing medium to assist online learners and take recommender systems to the next level.

The purpose of this thesis is to research how consumer recommendations can be used in terms of marketing and sales purposes in the Fast Moving Consumer Goods sector. Due to technological changes and a shift from traditional to digital marketing, the way of communicating to the audience has changed. Consumer recommendations, micro influencers and social media platforms are a part of digital marketing that is growing progressively. The paradigm of the thesis is interpretivism with following a qualitative research method, inductive reasoning and a case study in focus. The findings show that digital marketing, thanks to aspects as social media and e-commerce, has been growing, however has not completely taken over the marketing field. Influencers are getting a bigger role as marketing tools with the shift of power evolving. Due to these reasons is why and how consumer recommendations have become such a big trend nowadays.

[EN] The use of artificially intelligent techniques to overcome specific shortcomings within e-learning systems is a well-researched area that keeps on evolving in an attempt to optimise such resourceful practices. The lack of personalization and the sentiment of isolation coupled with a feeling of being treated like all others, tends to discourage and push learners away from courses that are very well prepared academically and excellently projected intellectually. The use of recommender systems to deliver relevant information in a timely manner that is specifically differentiated to a unique learner is once more being investigated to alievate the e-learning issue of being impersonal. The application of such a technique also assists the learner by reducing information overload and providing learning material that can be shared, criticized and reviewed at one’s own pace. In this paper we propose the use of a fully automated recommender system based on recent AI developments together with Web 2.0 applications and socially networked technologies. We argue that such technologies have provided the extra capabilities that were required to deliver a realistic and practical interfacing medium to assist online learners and take recommender systems to the next level.

[EN] The use of artificially intelligent techniques to overcome specific shortcomings within e-learning systems is a well-researched area that keeps on evolving in an attempt to optimise such resourceful practices. The lack of personalization and the sentiment of isolation coupled with a feeling of being treated like all others, tends to discourage and push learners away from courses that are very well prepared academically and excellently projected intellectually. The use of recommender systems to deliver relevant information in a timely manner that is specifically differentiated to a unique learner is once more being investigated to alievate the e-learning issue of being impersonal. The application of such a technique also assists the learner by reducing information overload and providing learning material that can be shared, criticized and reviewed at one’s own pace. In this paper we propose the use of a fully automated recommender system based on recent AI developments together with Web 2.0 applications and socially networked technologies. We argue that such technologies have provided the extra capabilities that were required to deliver a realistic and practical interfacing medium to assist online learners and take recommender systems to the next level.

Online shopping cart abandonment continues to be a widely studied phenomenon in the e-commerce space. While many existing studies are designed to analyze a breadth of factors, our specialized study aims to further our understanding of purchase hesitation and subsequent cart abandonment specifically caused by shipping fees. In order to explore shipping fee-related purchase hesitation, we employed a two part study in which we 1) collected data regarding the amount of additional fees a user is willing to pay at different price points as well as the user's psychological responses to these fees, then we 2) integrated crowdsourcing techniques into the results of part 1 and proposed a new interaction model, CrowdShop, that aims to crowdsource users' orders together based on geo-location proximity in order to reach a minimum free shipping threshold. Furthermore, we presented CrowdShop as a functioning prototype to gauge user response and sentiment. In both parts of the study, we employed a grounded-theory approach along with statistical methods for the analysis of the data. In order to reach a free shipping threshold for online orders, we found that many users combine purchasing efforts with closely-located family members and friends. However, user success of reaching a minimum order amount was hindered by the manual nature of such a collaboration. Through user testing sessions, we received positive responses in favor of Crowdshop's ability to streamline and automate collaboration efforts. However, there still exists opportunities for further development regarding user concerns with CrowdShop.

[EN] The use of artificially intelligent techniques to overcome specific shortcomings within e-learning systems is a well-researched area that keeps on evolving in an attempt to optimise such resourceful practices. The lack of personalization and the sentiment of isolation coupled with a feeling of being treated like all others, tends to discourage and push learners away from courses that are very well prepared academically and excellently projected intellectually. The use of recommender systems to deliver relevant information in a timely manner that is specifically differentiated to a unique learner is once more being investigated to alievate the e-learning issue of being impersonal. The application of such a technique also assists the learner by reducing information overload and providing learning material that can be shared, criticized and reviewed at one’s own pace. In this paper we propose the use of a fully automated recommender system based on recent AI developments together with Web 2.0 applications and socially networked technologies. We argue that such technologies have provided the extra capabilities that were required to deliver a realistic and practical interfacing medium to assist online learners and take recommender systems to the next level.

This thesis lays a foundation for understanding people's perceptions and engagement in news production, during a time of limited transparency about why, how and where information spreads online. Through creating, operating, and examining a news organisation purpose-built for this project, it was possible to describe critical gains and losses for the practice and business of doing news online. In interpretation, the dissertation documents a process of social capital being built among participants in the news operation, and posits that this will translate into a social infrastructure that can sustain such operations.

Despite the significant amount of energy spent on Heating, Ventilation, and Air Conditioning (HVAC) at universities, thermal comfort conditions in campus buildings are frequently poor. Conventional HVAC management systems at universities are typically out of the hands of building occupants and facilities management departments have limited resources to involve them. These factors can lead to over-heating or over-cooling and undiagnosed mechanical issues. Previous research has shown that thermal comfort feedback, or participatory thermal sensing, can simultaneously improve energy efficiency and occupant comfort in university buildings. However, these studies have been limited to single campus buildings and restricted populations of occupants. The success, scalability, and sustainability of any participatory thermal sensing program is dependent upon ongoing participation that is meaningful to occupants and useful to facilities management. Therefore, research is warranted to explore patterns of voluntary participation, thermal comfort, and occupant satisfaction in extended, campus-wide deployments of participatory thermal sensing programs. The present research begins to address these gaps in the context of an 18-month, campus-wide deployment of TherMOOstat, a participatory thermal sensing app and HVAC management system at University of California, Davis.

Despite the significant amount of energy spent on Heating, Ventilation, and Air Conditioning (HVAC) at universities, thermal comfort conditions in campus buildings are frequently poor. Conventional HVAC management systems at universities are typically out of the hands of building occupants and facilities management departments have limited resources to involve them. These factors can lead to over-heating or over-cooling and undiagnosed mechanical issues. Previous research has shown that thermal comfort feedback, or participatory thermal sensing, can simultaneously improve energy efficiency and occupant comfort in university buildings. However, these studies have been limited to single campus buildings and restricted populations of occupants. The success, scalability, and sustainability of any participatory thermal sensing program is dependent upon ongoing participation that is meaningful to occupants and useful to facilities management. Therefore, research is warranted to explore patterns of voluntary participation, thermal comfort, and occupant satisfaction in extended, campus-wide deployments of participatory thermal sensing programs. The present research begins to address these gaps in the context of an 18-month, campus-wide deployment of TherMOOstat, a participatory thermal sensing app and HVAC management system at University of California, Davis.

Identifying accurate biomarkers of cognitive decline is essential for advancing early diagnosis and prevention therapies in Alzheimer's disease. The Alzheimer's disease DREAM Challenge was designed as a computational crowdsourced project to benchmark the current state-of-the-art in predicting cognitive outcomes in Alzheimer's disease based on high dimensional, publicly available genetic and structural imaging data. This meta-analysis failed to identify a meaningful predictor developed from either data modality, suggesting that alternate approaches should be considered for prediction of cognitive performance.

Crowdsourcing is a powerful tool to harness citizen assessments in some complex decision tasks. When multiple annotators provide their individual labels a more reliable collective decision is obtained if the individual reliability parameters are incorporated in the decision making procedure. The well-known Maximum A Posteriori (MAP) rule weights the individual labels in proportion to the annotators’ reliability. In this work we analyze how the crowdsourcing system performance is degraded with the use of noisy annotators’ reliability parameters and we derive an alternative MAP based rule to be applied when these parameters are neither known nor even estimated by the decision system. We also derive analytical expected error rates and their upper bounds obtained by each rule as a useful tool to estimate the number of necessary annotators in the collective decision system depending on the level of noise present in the estimated reliability parameters., Peer Reviewed, Postprint (published version)

In light of the massive growth of creation and consumption of video, in this thesis I explore the concept of user-enhanced video browsing and evaluate the quantitative and qualitative effects of this approach. For this purpose, I create a unique prototype based on the VLC media player. The player interprets user-generated video tagging and annotations in my designed format, and allows the viewing of multiple event layers to create a personalized video playback. I perform evaluations in two user studies for this work. Among my observations I find benefits in navigation, personalization, consumption, and comprehension. I then look at the way viewers behave when contributing data of annotations and tagging. I find their preferred tasks, their perceived quality of other contributions, as well as their opinions on this system. I conclude with a discussion of the results and list possible use-cases for this concept.

In light of the massive growth of creation and consumption of video, in this thesis I explore the concept of user-enhanced video browsing and evaluate the quantitative and qualitative effects of this approach. For this purpose, I create a unique prototype based on the VLC media player. The player interprets user-generated video tagging and annotations in my designed format, and allows the viewing of multiple event layers to create a personalized video playback. I perform evaluations in two user studies for this work. Among my observations I find benefits in navigation, personalization, consumption, and comprehension. I then look at the way viewers behave when contributing data of annotations and tagging. I find their preferred tasks, their perceived quality of other contributions, as well as their opinions on this system. I conclude with a discussion of the results and list possible use-cases for this concept.

The research focuses on guiding the affected population towards a safe location in a disaster area by utilizing their self-help capacity with prevalent mobile technology. In contrast to the traditional centralized information management systems for disaster response, this research proposes a decen- tralized computer-supported management system in which affected people can lead themselves to safety and simultaneously serve as field sensors that share information about the disaster situation. A literature study together with contextual inquiries (field observations) were first conducted to analyse and identify existing problematic areas in order to envision a feasible, efficient and effective system. Based on literature from the field of disaster sociology and humanitarian operation experience, it was clear that the affected people in a disaster are not simply helpless victims; instead, they consist of capable human beings who tend to act rationally and proactively in a united and helpful manner. In times of collective stress, they are even able to creatively use whatever means of technology are at hand. This group forms the majority of the population that is affected by a disaster. They are distributed over the disaster area and witness the disaster first hand. As a result, these individuals form a massive potential resource for collecting first hand knowledge about the disaster. Field observations indicated that tailored mobile navigation support is lacking. Likewise, situation map-map- ping support has appeared to be limited, fragmented, and funnelled to one “plotter”. The literature and field observations showed that there is a need for (1) navigation support of the affected people, (2) collaborative map-making support, and (3) the integration of these support functions into one informa- tion system to establish a harmonized, effective, and safe, disaster response. As part of the research, prototypes of these two support functions and the integrated system were de, Mediamatics, Electrical Engineering, Mathematics and Computer Science

The research focuses on guiding the affected population towards a safe location in a disaster area by utilizing their self-help capacity with prevalent mobile technology. In contrast to the traditional centralized information management systems for disaster response, this research proposes a decen- tralized computer-supported management system in which affected people can lead themselves to safety and simultaneously serve as field sensors that share information about the disaster situation. A literature study together with contextual inquiries (field observations) were first conducted to analyse and identify existing problematic areas in order to envision a feasible, efficient and effective system. Based on literature from the field of disaster sociology and humanitarian operation experience, it was clear that the affected people in a disaster are not simply helpless victims; instead, they consist of capable human beings who tend to act rationally and proactively in a united and helpful manner. In times of collective stress, they are even able to creatively use whatever means of technology are at hand. This group forms the majority of the population that is affected by a disaster. They are distributed over the disaster area and witness the disaster first hand. As a result, these individuals form a massive potential resource for collecting first hand knowledge about the disaster. Field observations indicated that tailored mobile navigation support is lacking. Likewise, situation map-map- ping support has appeared to be limited, fragmented, and funnelled to one “plotter”. The literature and field observations showed that there is a need for (1) navigation support of the affected people, (2) collaborative map-making support, and (3) the integration of these support functions into one informa- tion system to establish a harmonized, effective, and safe, disaster response. As part of the research, prototypes of these two support functions and the integrated system were de, Mediamatics, Electrical Engineering, Mathematics and Computer Science

Táto diplomová práca sa zaoberá vhodným výberom meracieho systému pre vyhodnocovanie dlhodobého merania vlastností mobilných sietí LTE. Práca popisuje mobilnú sieť LTE, architektúru a kvalitu služby v LTE. V ďalšej časti sú preštudované rôzne známe merania tejto siete. Následne je predstavené riešenie vo forme dlhodobého merania pomocou mobilnej aplikácie Open-RMBT, ktorá ukladá svoje výsledky do databázy na zvolený server. Ďalšia časť je zameraná na zhrnutie výberu mobilného telefónu určeného na meranie, spustenie pilotného merania a následne uskutočnené dlhodobé meranie. V poslednej časti je predstavená interaktívna aplikácia na zobrazovanie grafov a následne vyhodnotené výsledky merania. Výsledky práce sú zhrnuté v závere., The aim of this thesis is to choose the proper measurement system for evaluating the performance of the LTE mobile networks based on long-term measurements. The thesis describes mobile network LTE, its architecture and introducts the Quality of Service in the LTE. In the next part are studied the known measurement methods and furthermore, there is proposed a solution in the form of long-term measurement. This measurement uses the mobile phone application Open-RMBT, which stores the measurement results into the database on the chosen server. The following part focuses on the issue of choosing a suitable mobile device for the measurement, then it describes a pilot testing and the followed run of the long-term measurements. In the last part is introduced the interractive application to plot graphs. Measurement results are discussed in the conclusion.

Despite the Department of Defense’s (DOD) investment in programs designed to advance sociocultural knowledge, the DOD lacks a shared repository through which all entities can aggregate, visualize, and share sociocultural data across the enterprise. A gap analysis of DOD’s desired and actual ability to achieve a sociocultural understanding reveals three shortcomings: data, repository, and collaboration. Therefore, we created a proof of concept that bridges the sociocultural gap by harnessing the overlooked potential of deployed service members and their cross-cultural experiences. Service member observations form an untapped resource of sociocultural data; this existing wellspring of sociocultural information needs to be collected and indexed using a common framework. Residing in a geodatabase and interfaced via a crowdsourced Geographic Information System (GIS), this framework aggregates the collected data of service member narratives for the greater Joint Force, thereby creating a dynamic and collaborative living repository. Combining an anthropologically sound and operationally relevant framework with the capabilities of GIS results in a solution that will allow DOD personnel to populate, visualize, and share near-real-time cultural data relevant to military operations across all services. This DOD enterprise solution can enhance the nation’s armed forces’ strategic performance through the application of culturally adept military power., http://archive.org/details/socioculturalgeo1094545806, Sergeant First Class, United States Army, Major, United States Army, Approved for public release; distribution is unlimited.

Táto diplomová práca sa zaoberá vhodným výberom meracieho systému pre vyhodnocovanie dlhodobého merania vlastností mobilných sietí LTE. Práca popisuje mobilnú sieť LTE, architektúru a kvalitu služby v LTE. V ďalšej časti sú preštudované rôzne známe merania tejto siete. Následne je predstavené riešenie vo forme dlhodobého merania pomocou mobilnej aplikácie Open-RMBT, ktorá ukladá svoje výsledky do databázy na zvolený server. Ďalšia časť je zameraná na zhrnutie výberu mobilného telefónu určeného na meranie, spustenie pilotného merania a následne uskutočnené dlhodobé meranie. V poslednej časti je predstavená interaktívna aplikácia na zobrazovanie grafov a následne vyhodnotené výsledky merania. Výsledky práce sú zhrnuté v závere., The aim of this thesis is to choose the proper measurement system for evaluating the performance of the LTE mobile networks based on long-term measurements. The thesis describes mobile network LTE, its architecture and introducts the Quality of Service in the LTE. In the next part are studied the known measurement methods and furthermore, there is proposed a solution in the form of long-term measurement. This measurement uses the mobile phone application Open-RMBT, which stores the measurement results into the database on the chosen server. The following part focuses on the issue of choosing a suitable mobile device for the measurement, then it describes a pilot testing and the followed run of the long-term measurements. In the last part is introduced the interractive application to plot graphs. Measurement results are discussed in the conclusion.

Táto diplomová práca sa zaoberá vhodným výberom meracieho systému pre vyhodnocovanie dlhodobého merania vlastností mobilných sietí LTE. Práca popisuje mobilnú sieť LTE, architektúru a kvalitu služby v LTE. V ďalšej časti sú preštudované rôzne známe merania tejto siete. Následne je predstavené riešenie vo forme dlhodobého merania pomocou mobilnej aplikácie Open-RMBT, ktorá ukladá svoje výsledky do databázy na zvolený server. Ďalšia časť je zameraná na zhrnutie výberu mobilného telefónu určeného na meranie, spustenie pilotného merania a následne uskutočnené dlhodobé meranie. V poslednej časti je predstavená interaktívna aplikácia na zobrazovanie grafov a následne vyhodnotené výsledky merania. Výsledky práce sú zhrnuté v závere., The aim of this thesis is to choose the proper measurement system for evaluating the performance of the LTE mobile networks based on long-term measurements. The thesis describes mobile network LTE, its architecture and introducts the Quality of Service in the LTE. In the next part are studied the known measurement methods and furthermore, there is proposed a solution in the form of long-term measurement. This measurement uses the mobile phone application Open-RMBT, which stores the measurement results into the database on the chosen server. The following part focuses on the issue of choosing a suitable mobile device for the measurement, then it describes a pilot testing and the followed run of the long-term measurements. In the last part is introduced the interractive application to plot graphs. Measurement results are discussed in the conclusion.

U.S. officials have acknowledged that attackers of the lone-wolf and isolated-cell organizational type are on the rise and now pose a greater threat than major coordinated actions. Traditional intelligence methods, using a top-down approach with an emphasis on signals intelligence, are ill-equipped to identify and prevent terrorists using lone-wolf tactics. Crowdsourcing, as a problem-solving technique, is a relatively new idea but has shown great promise in tackling issues similar to the identification of lone-wolf terrorists. At its core, crowdsourcing is a method for thousands or even millions of people to contribute their knowledge, expertise, or skills towards a unified task. Done correctly, it has produced results unachievable by traditional tasking of humans or computers. This thesis identifies how the signals surrounding lone-wolf attacks are different and more subtle in nature from those mounted by organized terror groups. In turn, the thesis examines the potential benefits of crowdsourcing intelligence in order to strengthen the U.S. intelligence community’s ability to approach this emerging problem of lone-wolf terrorism. In short, this thesis proposes that the U.S. intelligence community harness the power of U.S. citizens to help prevent identify the subtle indictors presented by lone-wolf terrorists in order to prevent lone-wolf terrorist attacks., http://archive.org/details/crowdsourcingint1094545174, Outstanding Thesis, Lieutenant Commander, United States Navy, Approved for public release; distribution is unlimited.