Text classification using compression-based dissimilarity measures

Arguably, the most difficult task in text classification is to choose an appropriate set of features that allows machine learning algorithms to provide accurate classification. Most state-of-the-art techniques for this task involve careful feature engineering and a pre-processing stage, which may be too expensive in the emerging context of massive collections of electronic texts. In this paper, we propose efficient methods for text classification based on information-theoretic dissimilarity measures, which are used to define dissimilarity-based representations. These methods dispense with any feature design or engineering, by mapping texts into a feature space using universal dissimilarity measures; in this space, classical classifiers (e.g. nearest neighbor or support vector machines) can then be used. The reported experimental evaluation of the proposed methods, on sentiment polarity analysis and authorship attribution problems, reveals that it approximates, sometimes even outperforms previous state-of-the-art techniques, despite being much simpler, in the sense that they do not require any text pre-processing or feature engineering.

Tradutor da língua gestual portuguesa modelo de tradução bidireccional

O projeto tem como objetivo desenvolver e avaliar um modelo que facilita o acesso para pessoas surdas ou com deficiência auditiva, o acesso ao conteúdo digital - em particular o conteúdo educacional e objetos de aprendizagem – a criação de condições para uma maior inclusão social de surdos e deficientes auditivos. Pretende-se criar um modelo bidirecional, em que permite a pessoas com deficiências auditivas, possam se comunicar com outras pessoas, com a tradução da Língua Gestual Portuguesa (LGP) para a Língua Portuguesa (LP) e que outras pessoas não portadoras de qualquer deficiência auditiva possam por sua vez comunicar com os surdos ou deficientes auditivos através da tradução da LP para a LGP. Há um conjunto de técnicas que poderíamos nos apoiar para desenvolver o modelo e implementar a API de tradução da LGP em LP. Muitos estudos são feitos com base nos modelos escondidos de Markov (HMM) para efetuar o reconhecimento. Recentemente os estudos estão a caminhar para o uso de técnicas como o “Dynamic Time Warping” (DTW), que tem tido mais sucesso do que outras técnicas em termos de performance e de precisão. Neste projeto optamos por desenvolver a API e o Modelo, com base na técnica de aprendizagem Support Vector Machines (SVM) por ser uma técnica simples de implementar e com bons resultados demonstrados em reconhecimento de padrões. Os resultados obtidos utilizando esta técnica de aprendizagem foram bastante ótimos, como iremos descrever no decorrer do capítulo 4, mesmo sabendo que utilizamos dois dispositivos para capturar dados de descrição de cada gesto. Toda esta tese integra-se no âmbito do projeto científico/ investigação a decorrer no grupo de investigação GILT, sob a coordenação da professora Paula Escudeiro e suportado pela Fundação para Ciência e Tecnologia (FCT).

Structured Light System Calibration for Perception in Underwater Tanks

The process of visually exploring underwater environments is still a complex problem. Underwater vision systems require complementary means of sensor information to help overcome water disturbances. This work proposes the development of calibration methods for a structured light based system consisting on a camera and a laser with a line beam. Two different calibration procedures that require only two images from different viewpoints were developed and tested in dry and underwater environments. Results obtained show, an accurate calibration for the camera/projector pair with errors close to 1 mm even in the presence of a small stereos baseline.

Time series morphological analysis applied to biomedical signals events detection

Dissertation submitted in the fufillment of the requirements for the Degree of Master in Biomedical Engineering

Automatic eye blink detection using consumer web cameras

This research aims to advance blinking detection in the context of work activity. Rather than patients having to attend a clinic, blinking videos can be acquired in a work environment, and further automatically analyzed. Therefore, this paper presents a methodology to perform the automatic detection of eye blink using consumer videos acquired with low-cost web cameras. This methodology includes the detection of the face and eyes of the recorded person, and then it analyzes the low-level features of the eye region to create a quantitative vector. Finally, this vector is classified into one of the two categories considered —open and closed eyes— by using machine learning algorithms. The effectiveness of the proposed methodology was demonstrated since it provides unbiased results with classification errors under 5%

Hochdurchsatz-Screeningsystem zur Objekterkennung in Mikroskop-Farbbildern im Rahmen der Analyse pflanzlicher Pathogenresistenz

Color image processing, pattern recognition, machine vision, application

Modeling sequencing errors by combining Hidden Markov models.

Among the largest resources for biological sequence data is the large amount of expressed sequence tags (ESTs) available in public and proprietary databases. ESTs provide information on transcripts but for technical reasons they often contain sequencing errors. Therefore, when analyzing EST sequences computationally, such errors must be taken into account. Earlier attempts to model error prone coding regions have shown good performance in detecting and predicting these while correcting sequencing errors using codon usage frequencies. In the research presented here, we improve the detection of translation start and stop sites by integrating a more complex mRNA model with codon usage bias based error correction into one hidden Markov model (HMM), thus generalizing this error correction approach to more complex HMMs. We show that our method maintains the performance in detecting coding sequences.

Immune regulation and function of the NOD-like receptors NLRC5 and NLRP3

AbstractThe vertebrate immune system is composed of the innate and the adaptive branches. Innate immune cells represent the first line of defense and detect pathogens through pattern recognition receptors (PRRs), detecting evolutionary conserved pathogen- and danger- associated molecular patterns. Engagement of these receptors initiates the inflammatory response, but also instructs antigen-specific adaptive immune cells. NOD-like receptors (NLRs) are an important group of PRRs, leading to the production of inflammatory mediators and favoring antigen presentation to Τ lymphocytes through the regulation of major histocompatibility complex (MHC) molecules.In this work we focused our attention on selected NOD-like receptors (NLRs) and their role at the interface between innate and adaptive immunity. First, we describe a new regulatory mechanism controlling IL-1 production. Our results indicate that type I interferons (IFNs) block NLRP1 and NLRP3 inflammasome activity and interfere with LPS-driven proIL-Ια and -β induction. As type I IFNs are produced upon viral infections, these anti-inflammatory effects of type I IFN could be relevant in the context of superinfections, but could also help explaining the efficacy of IFN-β in multiple sclerosis treatment.The second project addresses the role of a novel NLR family member, called NLRC5. The function of this NLR is still matter of debate, as it has been proposed as both an inhibitor and an activator of different inflammatory pathways. We found that the expression of this protein is restricted to immune cells and is positively regulated by IFNs. We generated Nlrc5-deficient mice and found that this NLR plays an essential role in Τ, NKT and, NK lymphocytes, in which it drives the expression of MHC class I molecules. Accordingly, we could show that CD8+ Τ cell-mediated killing of target lymphocytes lacking NLRC5 is strongly impaired. Moreover, NLRC5 expression was found to be low in many lymphoid- derived tumor cell lines, a mechanism that could be exploited by tumors to escape immunosurveillance.Finally, we found NLRC5 to be involved in the production of IL-10 by CD4+ Τ cells, as Nlrc5- deficient Τ lymphocytes produced less of this cytokine upon TCR triggering. In line with these observations, Mrc5-deficient CD4+ Τ cells expanded more than control cells when transferred into lymphopenic hosts and led to a more rapid appearance of colitis symptoms. Therefore, our work gives novel insights on the function of NLRC5 by using knockout mice, and strongly supports the idea that NLRs direct not only innate, but also adaptive immune responses.

Automatic Mallampati Classification Using Active Appearance Models

Difficult tracheal intubation assessment is an important research topic in anesthesia as failed intubations are important causes of mortality in anesthetic practice. The modified Mallampati score is widely used, alone or in conjunction with other criteria, to predict the difficulty of intubation. This work presents an automatic method to assess the modified Mallampati score from an image of a patient with the mouth wide open. For this purpose we propose an active appearance models (AAM) based method and use linear support vector machines (SVM) to select a subset of relevant features obtained using the AAM. This feature selection step proves to be essential as it improves drastically the performance of classification, which is obtained using SVM with RBF kernel and majority voting. We test our method on images of 100 patients undergoing elective surgery and achieve 97.9% accuracy in the leave-one-out crossvalidation test and provide a key element to an automatic difficult intubation assessment system.

Ecdysone triggered PGRP-LC expression controls Drosophila innate immunity.

Throughout the animal kingdom, steroid hormones have been implicated in the defense against microbial infection, but how these systemic signals control immunity is unclear. Here, we show that the steroid hormone ecdysone controls the expression of the pattern recognition receptor PGRP-LC in Drosophila, thereby tightly regulating innate immune recognition and defense against bacterial infection. We identify a group of steroid-regulated transcription factors as well as two GATA transcription factors that act as repressors and activators of the immune response and are required for the proper hormonal control of PGRP-LC expression. Together, our results demonstrate that Drosophila use complex mechanisms to modulate innate immune responses, and identify a transcriptional hierarchy that integrates steroid signalling and immunity in animals.

Can MMTV exploit TLR4?

The recognition of microbial pathogens based on their molecular patterns is essential for host defense. Recently, Toll-like receptors have been shown not only to recognize viruses as well as bacteria and fungi, but also to trigger an efficient immune response. A recent publication proposed that the retrovirus mouse mammary tumor virus exploits the pattern-recognition receptor Toll-like receptor 4 to achieve more efficient infection.

The inflammasomes.

Inflammasomes are molecular platforms activated upon cellular infection or stress that trigger the maturation of proinflammatory cytokines such as interleukin-1beta to engage innate immune defenses. Strong associations between dysregulated inflammasome activity and human heritable and acquired inflammatory diseases highlight the importance this pathway in tailoring immune responses. Here, we comprehensively review mechanisms directing normal inflammasome function and its dysregulation in disease. Agonists and activation mechanisms of the NLRP1, NLRP3, IPAF, and AIM2 inflammasomes are discussed. Regulatory mechanisms that potentiate or limit inflammasome activation are examined, as well as emerging links between the inflammasome and pyroptosis and autophagy.

Host innate immune responses to microbial pathogens.

Sepsis is among the leading causes of death worldwide and its incidence is increasing. Defined as the host response to infection, sepsis is a clinical syndrome considered to be the expression of a dysregulated immune reaction induced by danger signals that may lead to organ failure and death. Remarkable progresses have been made in our understanding of the molecular basis of host defenses in recent years. The host defense response is initiated by innate immune sensors of danger signals designated under the collective name of pattern-recognition receptors. Members of the family of microbial sensors include the complement system, the Toll-like receptors, the nucleotide-binding oligomerization domainlike receptors, the RIG-I-like helicases and the C-type lectin receptors. Ligand-activated pattern-recognition receptors kick off a cascade of intracellular events resulting in the expression of co-stimulatory molecules and release of effector molecules playing a fundamental role in the initiation of the innate and adaptive immune responses. Fine tuning of proinflammatory and anti-inflammatory reactions is critical for keeping the innate immune response in check. Overwhelming or dysregulated responses induced by infectious stimuli may have dramatic consequences for the host as shown by the profound derangements observed in sepsis. Unfortunately, translational research approaches aimed at the development of therapies targeting newly identified innate immune pathways have not held their promises. Indeed, all recent clinical investigations of adjunctive anti-sepsis treatments had little, if any, impact on morbidity and all-cause mortality of sepsis. Dissecting the mechanisms underlying the transition from infection to sepsis is essential for solving the sepsis enigma. Important components of the puzzle have already been identified, but the hunt must go on in the laboratory and at the bedside.

IL-13 induces expression of CD36 in human monocytes through PPARgamma activation.

The class B scavenger receptor CD36 is a component of the pattern recognition receptors on monocytes that recognizes a variety of molecules. CD36 expression in monocytes depends on exposure to soluble mediators. We demonstrate here that CD36 expression is induced in human monocytes following exposure to IL-13, a Th2 cytokine, via the peroxisome proliferator-activated receptor (PPAR)gamma pathway. Induction of CD36 protein was paralleled by an increase in CD36 mRNA. The PPARgamma pathway was demonstrated using transfection of a PPARgamma expression plasmid into the murine macrophage cell line RAW264.7, expressing very low levels of PPARgamma, and in peritoneal macrophages from PPARgamma-conditional null mice. We also show that CD36 induction by IL-13 via PPARgamma is dependent on phospholipase A2 activation and that IL-13 induces the production of endogenous 15-deoxy-Delta12,14-prostaglandin J2, an endogenous PPARgamma ligand, and its nuclear localization in human monocytes. Finally, we demonstrate that CD36 and PPARgamma are involved in IL-13-mediated phagocytosis of Plasmodium falciparum-parasitized erythrocytes. These results reveal a novel role for PPARgamma in the alternative activation of monocytes by IL-13, suggesting that endogenous PPARgamma ligands, produced by phospholipase A2 activation, could contribute to the biochemical and cellular functions of CD36.

Innate receptors for adaptive immunity.

Pattern recognition receptors (PRRs) are commonly known as sensor proteins crucial for the early detection of microbial or host-derived stress signals by innate immune cells. Interestingly, some PRRs are also expressed and functional in cells of the adaptive immune system. These receptors provide lymphocytes with innate sensing abilities; for example, B cells express Toll-like receptors, which are important for the humoral response. Strikingly, certain other NOD-like receptors are not only highly expressed in adaptive immune cells, but also exert functions related specifically to adaptive immune system pathways, such as regulating antigen presentation. In this review, we will focus particularly on the current understanding of PRR functions intrinsic to B and T lymphocytes; a developing aspect of PRR biology.