Publications

Over-The-Top Catch-up TV content-aware caching

Nogueira J., Gonzalez D., Guardalben L., Sargento S.

Proceedings - IEEE Symposium on Computers and Communications

2016

Conference Paper

Abstract:

The migration of popular Catch-up TV services to modern Over-The-Top (OTT) multimedia delivery infrastructures creates a wide set of scalability challenges which are commonly addressed using Content Delivery Networks (CDNs) relying on caching nodes close to users. The use of general-purpose caching nodes, tailored for generic web content, is far from optimal as it does not consider the particularities of Catch-up TV content, namely its dynamic popularity behavior, superstar effects, and relevance decay, as shown in existing scientific literature. Since caches are limited in size and are relatively small when compared to the whole catalog of available Catch-up TV content, which may contain tens of thousands of TV programs, it is crucial to make the most out of the available resources. To address these issues, this paper proposes a novel content-aware cache replacement algorithm, Most Popularly Used (MPU), capable of taking advantage of content demand forecasts built using machine learning models, to significantly outperform traditional cache replacement policies, such as Least Recently Used (LRU), Least Frequently Used (LFU), and First-In-First-Out (FIFO), and approach the optimal theoretical hit-ratio limits. MPU leverages millions of Catch-up TV request logs to validate its results under realistic conditions.

Overview of computer-assisted language learning for European Portuguese AT L 2F

Pellegrini T., Ling W., Silva A., Correia R., Trancoso I., Baptista J., Mamede N.

CSEDU 2012 - Proceedings of the 4th International Conference on Computer Supported Education

2012

Conference Paper

Abstract:

In this paper, we give an overview of our research in Computer-Assisted Language Learning for European Portuguese, to show how our long-time experience in spoken language processing allowed to propose multimedia documents as learning material. Web-based serious game were also introduced to cover aspects of listening, reading, and writing skills. One fundamental aspect of all our tools remains in the fully-automatic generation of the curriculum. This is very valuable for teachers, saving them time in search for motivating materials of appropriate quality, level and topic. A Web portal was recently created to make all our tools publicly available at http://call.l2f.inesc-id.pt/reap.public.

PACE your network: Fair and controllable multi-tenant data center networks

Carvalho T., Kim H.S., Neves N.

IEEE International Conference on Communications

2013

Conference Paper

Abstract:

Multi-tenant data centers host a high diversity of applications with continuously changing demands. Applications require response times ranging from a few microseconds to seconds. Therefore, network traffic within the data center needs to be managed in order to meet the requested SLAs. Current feedback congestion control protocols may be too slow to converge to a stable state under high congestion situations. Sudden bursts of traffic from heterogeneous sources may render any reactive control inefficient. In this paper, we propose PACE, a preventive explicit allocation congestion control protocol that controls resource allocations dynamically and efficiently. PACE specifically addresses Data Center requirements: efficient network usage, flow completion time guarantees, fairness in resource allocation, and scalability to hundreds of concurrent flows. PACE provides micro-allocation of network resources within dynamic periods, lossless communication, fine-grained prioritization of flows, and fast adaptation of allocations to the arrival of new flows. We simulate PACE and compare it with recent proposed protocols specially addressed to Data Centers. We demonstrate that PACE is fairer, in particular for short flows, flows with different RTTs and a higher number of concurrent flows. It also maintains high efficiency and controlled queue usage when exposed to sudden bursts.

Parallel combination of speech streams for improved ASR

Miranda J., Neto J.P., Black A.W.

13th Annual Conference of the International Speech Communication Association 2012, INTERSPEECH 2012

2012

Conference Paper

Abstract:

In a growing number of applications, such as simultaneous interpretation, audio or text may be available conveying the same information in different languages. Thesedifferent views contain redundant information that can be explored to enhance the performance of speech and language processing applications. We propose a method that directly integrates ASR word graphs or lattices and phrase tables from an SMT system to combine such parallel speech data and improve ASR performance. We apply this technique to speeches from four European Parliament committees and obtain a 16.6% relative improvement (20.8% after a second iteration) in WER, when Portuguese and Spanish interpreted versions are combined with the original English speeches. Our results indicate that further improvements may be possible by including additional languages. Index Terms: multistream combination, speech recognition, machine translation

Parameters that affect safety message delay in sparse infrastructure-less vehicular networks

Reis A.B., Sargento S., Tonguz O.K.

2014 IEEE International Conference on Communications, ICC 2014

2014

Conference Paper

Abstract:

In sparse highway vehicular networks, the high probability for network disconnection at the initial stages of introducing the DSRC technology can be mitigated by the deployment of fixed infrastructure points known as Road Side Units (RSU). However, due to the cost associated with the deployment and maintenance of significant numbers of RSUs, it is highly unlikely that the majority of highways will be seeing RSU support in the near future. In this paper we study the impact of specific vehicular network parameters in the communication delays in infrastructure-less highway scenarios: first, the deceleration of vehicles, and consequently, a decrease in their separation from succeeding vehicles; and second, the transmission power of the IEEE 802.11p radio, which can be increased to achieve faster connectivity with the succeeding vehicle. Our results show that the connectivity of sparse vehicular networks can be improved substantially by varying these parameters.

Paraphrasing 4 microblog normalization

Ling W., Dyer C., Black A.W., Trancoso I.

EMNLP 2013 - 2013 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference

2013

Conference Paper

Abstract:

Compared to the edited genres that have played a central role in NLP research, microblog texts use a more informal register with nonstandard lexical items, abbreviations, and free orthographic variation. When confronted with such input, conventional text analysis tools often perform poorly. Normalization — replacing orthographically or lexically idiosyncratic forms with more standard variants — can improve performance. We propose a method for learning normalization rules from machine translations of a parallel corpus of microblog messages. To validate the utility of our approach, we evaluate extrinsically, showing that normalizing English tweets and then translating improves translation quality (compared to translating unnormalized text) using three standard web translation services as well as a phrase-based translation system trained on parallel microblog data.

Partially Monotonic Learning for Neural Networks

Trindade J., Vinagre J., Fernandes K., Paiva N., Jorge A.

Advances in Intelligent Data Analysis XIX

2021

Book Chapter

Abstract:

In the past decade, we have witnessed the widespread adoption of Deep Neural Networks (DNNs) in several Machine Learning tasks. However, in many critical domains, such as healthcare, finance, or law enforcement, transparency is crucial. In particular, the lack of ability to conform with prior knowledge greatly affects the trustworthiness of predictive models. This paper contributes to the trustworthiness of DNNs by promoting monotonicity. We develop a multi-layer learning architecture that handles a subset of features in a dataset that, according to prior knowledge, have a monotonic relation with the response variable. We use two alternative approaches: (i) imposing constraints on the model’s parameters, and (ii) applying an additional component to the loss function that penalises non-monotonic gradients. Our method is evaluated on classification and regression tasks using two datasets. Our model is able to conform to known monotonic relations, improving trustworthiness in decision making, while simultaneously maintaining small and controllable degradation in predictive ability.

Patient innovation under rare diseases and chronic needs

Oliveira P., Zejnilovic L., Canhao H., von Hippel E.

Orphanet Journal of Rare Diseases

2014

Conference Paper

Abstract:

The patients afflicted by more than 7000 rare diseases are in “orphan” markets. They can expect little help from producers in the form of specialized products, and they have strong incentives to develop, or adopt solutions developed by peers, to help them cope with the diseases. We don’t know the extent to which patients and caregivers respond to these incentives and innovate, how they perceive impact of their solutions, and how they share their solutions with others. The objectives of this work are: to measure frequency of patient innovation in a population of rare diseases patients; to measure efforts by patients to share their solutions with others; to explore which factors drive patients to come-up with solutions and share them with others.

Peer Adoption and Development of Health Innovations by Patients: National Representative Study of 6204 Citizens

Oliveira P., Zejnilovic L., Azevedo S., Rodrigues A.M., Canhão H.

Journal of Medical Internet Research

2019

Article

Abstract:

There is growing evidence that many patients and caregivers innovate by developing new solutions to cope with their health disorders. Given the easy access to vast internet resources and peers globally, it is increasingly important to understand what may influence user innovation and its adoption in health for improving individual well-being and ensuring their safety, in particular, how interactions with peers and physicians or search behavior, along with sociodemographics, may influence the decision to develop a solution or adopt one developed by a peer.