A5089099115- Natural Language Processing Techniques
- Topic Modeling
- Parallel Computing and Optimization Techniques
- Web Data Mining and Analysis
- Interconnection Networks and Systems
- Advanced Text Analysis Techniques
- Sentiment Analysis and Opinion Mining
- Speech Recognition and Synthesis
- Biomedical Text Mining and Ontologies
- Embedded Systems Design Techniques
- Semantic Web and Ontologies
- Human Mobility and Location-Based Analysis
- Distributed and Parallel Computing Systems
- Urban Transport and Accessibility
- Voice and Speech Disorders
- Intellectual Property and Patents
- Text Readability and Simplification
- Neural Networks and Applications
- Academic integrity and plagiarism
- Mobile Crowdsensing and Crowdsourcing
- Spam and Phishing Detection
- Transportation Planning and Optimization
- Machine Learning in Materials Science
- Speech and dialogue systems
- Bipolar Disorder and Treatment
Universitat Pompeu Fabra
2007-2021
Barcelona Media
2010-2012
Intel (United States)
2009-2011
Universitat Politècnica de Catalunya
1991-2005
Laboratoire d'Optique Appliquée
1957
Epidemiological studies suggest that bipolar disorder has a prevalence of about 1% in European countries, becoming one the most disabling illnesses working age adults, and often long-term persistent with complex management treatment. Therefore, capacity home monitoring for patients this is crucial their quality life. The current paper introduces use speech-based information as an easy-to-record, ubiquitous non-intrusive health sensor suitable monitoring, its application framework on...
This paper presents a novel scheme to schedule loops for clustered microarchitectures. The is based on preliminary cluster assignment phase implemented through graph partitioning techniques followed by scheduling that integrates register allocation and spill code generation. shown be very effective due its global view of the whole while partition generated. Results show significant speedup when compared with previously proposed techniques. For some processor configuration average SPECfp95...
This paper presents AGAMOS, a technique to modulo schedule loops on clustered microarchitectures. The proposed scheme uses multilevel graph partitioning strategy distribute the workload among clusters and reduces number of intercluster communications at same time. Partitioning is guided by approximate schedules (i.e., pseudoschedules), which take into account all constraints that influence final schedule. To further reduce communications, heuristics for instruction replication are included....
This work presents a modulo scheduling framework for clustered ILP processors that integrates the cluster assignment, instruction and register allocation steps in single phase. unified approach is more effective than traditional approaches based on sequentially performing some (or all) of three steps, since it allows optimizing global code generation problem instead searching optimal solutions to each individual step. Besides, avoids iterative nature approaches, which require repeated...
Patients with chronic obstructive pulmonary disease (COPD) suffer from voice changes respect to the healthy population. However, two issues remain be studied: how long-term speech elements such as prosody are affected; and whether physical effort medication also affect of patients COPD, if so, an automatic speech-based detection system COPD measurements can influenced by these changes. The aim current study is address both issues. To this end, long read control groups was recorded, following...
Lexical collocations are idiosyncratic combinations of two syntactically bound lexical items (e.g., "heavy rain" or "take a step"). Understanding their degree compositionality and idiosyncrasy, as well underlying semantics, is crucial for language learners, lexicographers downstream NLP applications. In this paper, we perform an exhaustive analysis current models collocation understanding. We first construct dataset apparitions in context, categorized into 17 representative semantic...
The importance of collocations, i.e. idiosyncratic binary word cooccurrences in the context second language learning has been repeatedly emphasized by scholars working field. Some went even so far as to argue that “vocabulary is collocation learning” (Hausmann, 1984, p. 395). Empirical studies confirm this argumentation. They show “collocation density” learner corpora nearly same native corpora, use collocations learners common it speakers. At time, they also find error rate about 32%...
This work presents a new compilation technique that uses instruction replication in order to reduce the number of communications executed on clustered microarchitecture. For such architectures, need communicate values between clusters can result significant performance loss. Inter-cluster be reduced by selectively replicating an appropriate set instructions. However, must done carefully since it may also degrade due increased contention place processor resources. The proposed scheme is built...
This paper presents a new modulo scheduling algorithm for clustered microarchitectures. The main feature of the proposed scheme is that assignment instructions to clusters done by means graph partitioning algorithms are guided pseudo-scheduler. pseudo-scheduler simplified version full instruction scheduler and estimates key constraints would be encountered in final schedule. process bi-directional includes on-the-fly spill code generation. evaluated against previous approaches using SPECfp95...