École Polytechnique Fédérale de Lausanne
Infoscience - École polytechnique fédérale de LausanneNot a member yet
191401 research outputs found
Sort by
London 1890s Ordnance Survey Text Layer
No full textThis dataset contains a sample of 10,000 (3.5%) out of a total of 285,846 text sequences extracted from the 1891–1896 Map of London by the Ordnance Survey (OS).
The methodology used for the automated recognition, linking, and sequencing of the text is detailed in the article Recognizing and Sequencing Multi-word Texts in Maps Using an Attentive Pointer by M. Zou et al., 2025.
The map is drawn at a scale of five-feet to the mile (c.a. 1:1,056). The text on the map is an invaluable source of information about the Greater London in the late Victorian period. It includes the names of streets, squares, parks, watercourses and even some estates ('Poplars', 'The Grange', 'Arbutus Lodge'). In addition, the map contains many details of the function of buildings and economic activity, such as factories ('Sweet Factory', 'Crown Linoleum Works', 'Imperial Flour Mills', 'Lion Brewery'), warehouses or commercial infrastructure ('Warehouse', 'Jamaica Wharf', 'Rag Store'), offices ('Offices'), etc. The map also mentions public buildings such as schools ('School Boys, Girls & Infants', 'Sunday School'), hospitals or clinics ('St. Saviour's Union Infirmary', 'Beulah Spa Hydropathic Establishment', 'South Western Fever Hospital'), railway stations ('Clapham Station'), post offices, banks, police stations, etc. Other social venues are also mentioned, such as public houses, i.e. pubs ('P.H.'), clubs, casinos, and recreational areas (e.g. 'Cricket Ground'). Special attention is given to churches, with a regular count of the number of seats (e.g. 'Baptist Chapel Seats for 600').
In addition, the map provides details that can be of great interest in the study of everyday life in London at the end of the 19th century. For example, there are numerous mentions of 'Stables', 'Drinking Fountain'[s] (or simply 'Fn.') or 'Urinal'[s]. Fire protection infrastructure is highlighted, e.g. fire plugs ('F.P.') and fire alarms ('F.A.'). The map also includes information on elevation (e.g. '11·6') and flood levels (e.g. 'High Water Mark of Ordinary Tides').DHI0.1.
EPFL-Smart-Kitchen-30 Collected data
No full textUnderstanding behavior requires datasets that capture humans while carrying out complex tasks. The kitchen is an excellent environment for assessing human motor and cognitive function, as many complex actions are naturally exhibited in kitchens from chopping to cleaning. Here, we introduce the EPFL-Smart-Kitchen-30 dataset, collected in a noninvasive motion capture platform inside a kitchen environment. Nine static RGB-D cameras, inertial measurement units (IMUs) and one head-mounted HoloLens~2 headset were used to capture 3D hand, body, and eye movements. The EPFL-Smart-Kitchen-30 dataset is a multi-view action dataset with synchronized exocentric, egocentric, depth, IMUs, eye gaze, body and hand kinematics spanning 29.7 hours of 16 subjects cooking four different recipes. Action sequences were densely annotated with 33.78 action segments per minute. Leveraging this multi-modal dataset, we propose four benchmarks to advance behavior understanding and modeling through
1) a vision-language benchmark,
2) a semantic text-to-motion generation benchmark,
3) a multi-modal action recognition benchmark,
4) a pose-based action segmentation benchmark.
> ⚠️ 3D pose and action annotations can be found at https://zenodo.org/records/15551913UPAMATHISTNEUPHUMMEL
Smart School Multimodal Dataset and Challenges
Get PDFAs part of a research project aiming to explore the notion of ‘smart school’ (especially for STEM education) in Estonia, we are developing classrooms and schools that incorporate data gathering not only from digital traces, but also physical ones (through a variety of sensors). This workshop contribution describes briefly the setting and our initial efforts in setting up a classroom that is able to generate such a multimodal dataset. The paper also describes some of the most important challenges that we are facing as we setup the project and attempt to build up such dataset, focusing on the specifics of doing it in an everyday, authentic school setting. We believe these challenges provide a nice sample of those that the multimodal learning analytics (MMLA) community will have to face as it transitions from an emergent to a mainstream community of research and practice.SCI-STI-D
Editorial
Get PDFLearning is a complex, mostly invisible process that happens across spaces, occurring in the physical world but also increasingly in virtual worlds or web-based spaces. In order to explore what happens in such blended learning experience, there is a need for multiple data sources that bring evidence from these different spaces. The present proceedings bring together two workshops co-located at the Learning Analytics and Knowledge (LAK‘17) conference in Vancouver (Canada): the 2nd Cross-LAK and the 6th Multimodal Learning Analytics (MMLA) workshop. The two workshops tackled the analysis of this complexity, from complementary perspectives. Our aim is to promote dialogue and the alignment of these research efforts across both subcommunities. Moreover, this collaboration is the seed of a Special Interest Group (SIG) that will be part of the Society of Learning Analytics Research (SoLAR). The goal of this SIG will be to advance the understanding of the learning process no matter where and how it happens.SCI-STI-D
Method and apparatus for presenting to a user of a wearable apparatus additional information related to an audio scene
No full textA method for presenting to a user of a wearable audio device a modified audio scene together with additional information related to the audio scene, comprising: capturing audio signals with a plurality of microphones; outputting an audio signal with a plurality of acoustical transducers; processing the captured audio signals, the processing comprising filtering, equalization, echoes processing and/or beamforming; separating audio sources from the processed audio signals; selecting at least one separated audio source; classifying at least one said selected audio source; retrieving additional information related to the classified audio source; presenting the additional information to the user.AVP-R-TTOLCA
Discovery and characterization of novel stable tau oligomeric complexes: Implications for the role of Tau/phospholipid interactions in regulating its functions in health and disease
No full textThe microtubule-associated protein Tau plays a central role in the pathogenesis of Alzheimer's disease. Although Tau interaction with membranes is thought to affect some of its physiological functions and its aggregation properties, the sequence determinants and the structural and functional consequences of such interactions remain poorly understood. Here, we report that the interaction of Tau with vesicles results in the formation of highly stable protein/phospholipid complexes. These complexes are toxic to primary hippocampal cultures and are detected by MC-1, an antibody recognizing pathological Tau conformations. The core of these complexes is comprised of the PHF6* and PHF6 hexapeptide motifs, the latter in a β-strand conformation. Studies using Tau-derived peptides enabled the design of mutants that disrupt Tau interactions with phospholipids without interfering with its ability to form fibrils, thus providing powerful tools for uncoupling these processes and investigating the role of membrane interactions in regulating Tau function, aggregation and toxicity.LMN
Detecting animals in African Savanna with UAVs and the crowds
Get PDFUnmanned aerial vehicles (UAVs) offer new opportunities for wildlife monitoring, with several advantages over traditional field-based methods. They have readily been used to count birds, marine mammals and large herbivores in different environments, tasks which are routinely performed through manual counting in large collections of images. In this paper, we propose a semi-automatic system able to detect large mammals in semi-arid Savanna. It relies on an animal-detection system based on machine learning, trained with crowd-sourced annotations provided by volunteers who manually interpreted sub-decimeter resolution color images. The system achieves a high recall rate and a human operator can then eliminate false detections with limited effort. Our system provides good perspectives for the development of data-driven management practices in wildlife conservation. It shows that the detection of large mammals in semi-arid Savanna can be approached by processing data provided by standard RGB cameras mounted on affordable fixed wings UAVs.LASI
Argument discovery via crowdsourcing
No full textThe amount of controversial issues being discussed on the Web has been growing dramatically. In articles, blogs, and wikis, people express their points of view in the form of arguments, i.e., claims that are supported by evidence. Discovery of arguments has a large potential for informing decision-making. However, argument discovery is hindered by the sheer amount of available Web data and its unstructured, free-text representation. The former calls for automatic text-mining approaches, whereas the latter implies a need for manual processing to extract the structure of arguments. In this paper, we propose a crowdsourcing-based approach to build a corpus of arguments, an argumentation base, thereby mediating the trade-off of automatic text-mining and manual processing in argument discovery. We develop an end-to-end process that minimizes the crowd cost while maximizing the quality of crowd answers by: (1) ranking argumentative texts, (2) pro-actively eliciting user input to extract arguments from these texts, and (3) aggregating heterogeneous crowd answers. Our experiments with real-world datasets highlight that our method discovers virtually all arguments in documents when processing only 25% of the text with more than 80% precision, using only 50% of the budget consumed by a baseline algorithm.LSI