INESC-ID

About: INESC-ID is a nonprofit organization based out in Lisbon, Portugal. It is known for research contribution in the topics: Field-programmable gate array & Control theory. The organization has 932 authors who have published 2618 publications receiving 37658 citations.

Influence-based abstraction for multiagent systems

Abstract: This paper presents a theoretical advance by which factored POSGs can be decomposed into local models. We formalize the interface between such local models as the influence agents can exert on one another; and we prove that this interface is sufficient for decoupling them. The resulting influence-based abstraction substantially generalizes previous work on exploiting weakly-coupled agent interaction structures. Therein lie several important contributions. First, our general formulation sheds new light on the theoretical relationships among previous approaches, and promotes future empirical comparisons that could come by extending them beyond the more specific problem contexts for which they were developed. More importantly, the influence-based approaches that we generalize have shown promising improvements in the scalability of planning for more restrictive models. Thus, our theoretical result here serves as the foundation for practical algorithms that we anticipate will bring similar improvements to more general planning contexts, and also into other domains such as approximate planning, decision-making in adversarial domains, and online learning.

The L2F Broadcast News Speech Recognition System

Abstract: Broadcast news play an important role in our lives provid-ing access to news, information and entertainment. The ex-istence of an automatic transcription is an important mediumthat not only can provide subtitles for inclusion of people withspecial needs or be an advantage on noisy and populated envi-ronments, but also because it enables data search and retrievecapabilities over the multimedia streams. In this work we willdescribe and evaluate the automatic speech recognition systemsdeveloped for two Iberian languages, European Portuguese andSpanish and also for Brazilian Portuguese, African Portugueseand English. The developed systems are fully automatic andcapable to subtitling in real-time Broadcast News stream with avery small delay.Index Terms: Speech Recognition, Broadcast News, Iberianlanguages, Accent, Online processing 1. Introduction The Broadcast News (BN) processing system developed at theSpoken Language Systems Lab of INESC-ID integrates sev-eral core technologies, in a pipeline architecture: jingle detec-tion, audio segmentation, automatic speech recognition, punc-tuation, capitalization, topic segmentation/indexation, summa-rization, and translation. The first modules of this system wereoptimized for on-line performance, given their deployment inthe fully automatic speech recognition subtitling system that isrunning on the main news shows of the public TV channel inPortugal (RTP), since March 2008.To our knowledge, the majority of subtitling systems de-scribed in the literature rely on speech-to-text alignment ratherthan full automatic speech recognition [1]. Re-speakers alsoare commonly used to simplify the original speech, and speechrecognition engines are adapted to the captioner voice [2].This paper concerns the third module in the pipeline -speech recognition, emphasizing the most recent improvements,and our efforts to port it to other languages (English and Span-ish), and to other varieties of Portuguese, namely those spokenin the South American and African continents.The development of a system for a new language is a chal-lenging task due to the need of new acoustic training data, vo-cabulary definition, lexicon generation and language model es-timation [3].The paper starts with a description of the main modules ofour recognition engine, emphasizing the two language indepen-dent components - feature extraction and decoder. The nextthree sections are devoted to the three varieties of Portuguesecovered by our system: the original one (European Portuguese,henceforth designated as EP), Brazilian Portuguese (BP), andAfrican Portuguese (AP). The porting efforts for the other twolanguages (European Spanish and American English) are de-scribed in Sections 6 and 7, respectively. For each of these sec-tions, we shall detail the corpora, vocabulary, and lexical andlanguage model generation, ending with performance results.The final section discusses the main advantages and shortcom-ings of these systems, namely in what concerns real time closecaptioning applications.

Approximate String Matching with Compressed Indexes

Abstract: A compressed full-text self-index for a text T is a data structure requiring reducedspace and able to search for patterns P in T . It can also reproduce any substring of T , thusactually replacing T . Despite the recent explosion of interest on compressed indexes, therehas not been much progress on functionalities beyond the basic exact search. In this paperwe focus on indexed approximate string matching (ASM), which is of great interest, say,in bioinformatics. We study ASM algorithms for Lempel-Ziv compressed indexes and forcompressed suffix trees/arrays. Most compressed self-indexes belong to one of these classes.We start by adapting the classical method of partitioning into exact search to self-indexes, andoptimize it over a representative of either class of self-index. Then, we show that a Lempel-Ziv index can be seen as an extension of the classical q -samples index. We give new insightson this type of index, which can be of independent interest, and then apply them to a Lempel-Ziv index. Finally, we improve hierarchical verification, a successful technique for sequentialsearching, so as to extend the matches of pattern pieces to the left or right. Most compressedsuffix trees/arrays support the required bidirectionality, thus enabling the implementation ofthe improved technique. In turn, the improved verification largely reduces the accesses to thetext, which are expensive in self-indexes. We show experimentally that our algorithms arecompetitive and provide useful space-time tradeoffs compared to classical indexes.

SleeveAR: Augmented Reality for Rehabilitation using Realtime Feedback

Abstract: We present an intelligent user interface that allows people to perform rehabilitation exercises by themselves under the offline supervision of a therapist. Every year, many people suffer injuries that require rehabilitation. This entails considerable time overheads since it requires people to perform specified exercises under the direct supervision of a therapist. Therefore it is desirable that patients continue performing exercises outside the clinic (for instance at home, thus without direct supervision), to complement in-clinic physical therapy. However, to perform rehabilitation tasks accurately, patients need appropriate feedback, as otherwise provided by a physical therapist, to ensure that these unsupervised exercises are correctly executed. Different approaches address this problem, providing feedback mechanisms to aid rehabilitation. Unfortunately, test subjects frequently report having trouble to completely understand the feedback thus provided, which makes it hard to correctly execute the prescribed movements. Worse, injuries may occur due to incorrect performance of the prescribed exercises, which severely hinders recovery. SleeveAR is a novel approach to provide real-time, active feedback, using multiple projection surfaces to provide effective visualizations. Empirical evaluation shows the effectiveness of our approach as compared to traditional video-based feedback. Our experimental results show that our intelligent UI can successfully guide subjects through an exercise prescribed (and demonstrated) by a physical therapist, with performance improvements between consecutive executions, a desirable goal to successful rehabilitation.

Mobility Network Evaluation in the User Perspective: Real-Time Sensing of Traffic Information in Twitter Messages

Abstract: We discuss a new type of mobility study focusing on subjective user opinions of mobility networks. Taking advantage of the growth in popularity of opinion mining in social media, and social media itself, we present the architecture of a system capable of automatically capturing user perspective toward a mobility network, based on web user-generated content. We discuss the value of acquiring this subjective information, especially for urban planners, and its contribution to an overall understanding of human mobility. We look at users as sensors of mobility dynamics, capable of providing an insight into the flaws of mobility networks. To achieve this, this chapter presents an initial attempt to use microblogging messages posted on Twitter (by users in transit) to perform real-time sensing of traffic-related information. We propose a text classification approach to the problem: we wish to automatically identify traffic-related messages posted on Twitter, among the millions of unrelated messages posted by users.