Search

Manual annotation of video segmentation datasets requires an immense amount of human effort, thus, reduction of human annotation costs is an active topic of research. While many papers deal with the propagation of masks through frames of a video, only a few results attempt to optimize annotation task selection. In this paper we present a deep learning based solution to the latter problem and train it using Reinforcement Learning. Our approach utilizes a modified version of the Dueling Deep Q-Network sharing weight parameters across the temporal axis of the video. This technique enables the trained agent to select annotation tasks from the whole video. We evaluate our annotation task selection method by means of a hierarchical supervoxel segmentation based mask propagation algorithm.

The goal of this paper is to review the relevant literature on autism questionnaires, models, analytic tools, and subgrouping, focusing on the opportunities and limitations. We examined how the size and nature of the database and the number and type of parameters measured determine the use of analytic tools and the expected type of results. To support our position about the examined aspects, we rely on various parts of the reviewed articles. We emphasize that the individual results in each article can only be interpreted in the light of the methods that brought about the conclusions. The reviewed literature suggests a heterogeneous palette in autism subtypes instead of distinct, well-characterized subgroups, or a single-dimensional continuous spectrum.

According to the recently proposed omnigenic theory, all expressed genes in a relevant tissue are contributing directly or indirectly to the manifestation of complex disorders such as autism. Thus, holistic approaches can be complementary in studying genetics of these complex disorders to focusing on a limited number of candidate genes. Gene interaction networks can be used for holistic studies of the omnigenic nature of autism. We used Louvain clustering on tissue-specific gene interaction networks and their subgraphs exclusively containing autism-related genes to study the effects of peripheral gene interactions. We observed that the autism gene clusters are significantly weaker connected to each other and the peripheral genes in non-neuronal tissues than in brain-related tissues. The biological functions of the brain clusters correlated well with previous findings on autism, such as synaptic signaling, regulation of DNA methylation, or regulation of lymphocyte activation, however, on the other tissues they did not enrich as significantly. Furthermore, ASD subjects with disruptive mutations in specific gene clusters show phenotypical differences compared to other disruptive variants carrying ASD individuals. Our results strengthen the omnigenic theory and can advance our understanding of the genetic background of autism.

Machine learning is making substantial progress in diverse applications. The success is mostly due to advances in deep learning. However, deep learning can make mistakes and its generalization abilities to new tasks are questionable. We ask when and how one can combine network outputs, when (i) details of the observations are evaluated by learned deep components and (ii) facts and rules are available. The Deep Consistence Seeking (DCS) machine seeks for consistent and deterministic event descriptions and improves the representation accordingly. The machine has an anomaly detection component that may trigger coherence seeking. Coherence seeking resolves conflicts between computational modules by preferring components with higher scores. We illustrate that context can help in correcting recognitions and in deriving training samples for self-training. We put these concepts into a general framework of cognition, by distinguishing creativity, rule extraction, verification, and symbol grounding. We demonstrate our approach in a driving scenario.

It has been long debated how the so called cognitive map, the set of place cells, develops in rat hippocampus. The function of this organ is of high relevance, since the hippocampus is the key component of the medial temporal lobe memory system, responsible for forming episodic memory, declarative memory, the memory for facts and rules that serve cognition in humans. Here, a general mechanism is put forth: We introduce the novel concept of Cartesian factors. We show a non-linear projection of observations to a discretized representation of a Cartesian factor in the presence of a representation of a complementing one. The computational model is demonstrated for place cells that we produce from the egocentric observations and the head direction signals. Head direction signals make the observed factor and sparse allothetic signals make the complementing Cartesian one. We present numerical results, connect the model to the neural substrate, and elaborate on the differences between this model and other ones, including Slow Feature Analysis [17] .

Stable unilamellar dipalmitoylphosphatidylcholine vesicles were produced by using oligo(malic acid) and cholesterol. Detailed physico-chemical characterization prove that by using oligo(malic acid) the substitution of PEGylated lipids for sterically stabilization comes possible. The polymer molecules cover the outer surface of spherical-shaped vesicles, and an asymmetrical composition occurs in the two leaflets of the phospholipid bilayer. The oligo(malic-acid) and cholesterol are enriched in the outer side assuring the stabilization of vesicles. Cholesterol plays an important role in the self-assembly of components as it makes the entering of oligomers possible deep into the polar head-region of lipids. The presence of oligo(malic acid) molecules does not induce degradation by hydrolysis of lipid molecules but the vesicle system turns into a sensitive form giving a possibility for pH sensitive targeting. Preliminary investigation on the investigated oligo(malic acid)-stabilized vesicles do not show any toxic effect promising their applicability in the field of liposomal drug delivery.

In this study we investigate the strategies of subjects in a complex divided attention task. We conducted a series of experiments with ten participants and evaluated their performance. After an extensive analysis, we identified four strategic measures that justify the achievement of the participants, by highlighting the individual differences and predicting performance in a regression analysis using generalized estimating equations. Selecting the more urgent task and user action between multiple simultaneous possibilities form two of the strategic decisions, respectively. The third one refers to choosing a response within the same task when the opportunity is present. The fourth and most important measure of strategy involves thinking ahead and executing an action before a situation would become critical. This latter one has the effect of reducing later cognitive load or timing constraints and it is shown to explain almost as much variance in performance as the other three, more straightforward predictors together. In addition to determining these strategic predictors, we also show how manipulating task difficulty induces a shift in strategy, thus impairing human performance in the rehearsed task. The results of this study indicate that considerable differences in the divided attention ability of normal subjects can be identified early and with simple measurements. The importance of describing and analyzing strategies is also emphasized, which can substantially influence performance in complex tasks and may serve training needs.

In monocular 3D human pose estimation a common setup is to first detect 2D positions and then lift the detection into 3D coordinates. Many algorithms suffer from overfitting to camera positions in the training set. We propose a siamese architecture that learns a rotation equivariant hidden representation to reduce the need for data augmentation. Our method is evaluated on multiple databases with different base networks and shows a consistent improvement of error metrics. It achieves state-of-the-art cross-camera error rate among algorithms that use estimated 2D joint coordinates only., Comment: Accepted to Neurocomputing

The existence of place cells (PCs), grid cells (GCs), border cells (BCs), and head direction cells (HCs) as well as the dependencies between them have been enigmatic. We make an effort to explain their nature by introducing the concept of Cartesian Factors. These factors have specific properties: (i) they assume and complement each other, like direction and position and (ii) they have localized discrete representations with predictive attractors enabling implicit metric-like computations. In our model, HCs make the distributed and local representation of direction. Predictive attractor dynamics on that network forms the Cartesian Factor “direction.” We embed these HCs and idiothetic visual information into a semi-supervised sparse autoencoding comparator structure that compresses its inputs and learns PCs, the distributed local and direction independent (allothetic) representation of the Cartesian Factor of global space. We use a supervised, information compressing predictive algorithm and form direction sensitive (oriented) GCs from the learned PCs by means of an attractor-like algorithm. Since the algorithm can continue the grid structure beyond the region of the PCs, i.e., beyond its learning domain, thus the GCs and the PCs together form our metric-like Cartesian Factors of space. We also stipulate that the same algorithm can produce BCs. Our algorithm applies (a) a bag representation that models the “what system” and (b) magnitude ordered place cell activities that model either the integrate-and-fire mechanism, or theta phase precession, or both. We relate the components of the algorithm to the entorhinal-hippocampal complex and to its working. The algorithm requires both spatial and lifetime sparsification that may gain support from the two-stage memory formation of this complex.

[This corrects the article DOI: 10.1371/journal.pone.0195131.].

We treat the problem of searching for hidden multi-dimensional independent auto-regressive processes (auto-regressive independent process analysis, AR-IPA). Independent subspace analysis (ISA) can be used to solve the AR-IPA task. The so-called separation theorem simplifies the ISA task considerably: the theorem enables one to reduce the task to one-dimensional blind source separation task followed by the grouping of the coordinates. However, the grouping of the coordinates still involves two types of combinatorial problems: (a) the number of the independent subspaces and their dimensions, and then (b) the permutation of the estimated coordinates are to be determined. Here, we generalize the separation theorem. We also show a non-combinatorial procedure, which—under certain conditions—can treat these two combinatorial problems. Numerical simulations have been conducted. We investigate problems that fulfill sufficient conditions of the theory and also others that do not. The success of the numerical simulations indicates that further generalizations of the separation theorem may be feasible.

Background The ATP-Binding Cassette (ABC)-transporter MultiDrug Resistance Protein 1 (MDR1) and Multidrug Resistance Related Protein 1 (MRP1) are expressed on the surface of enterocytes, which has led to the belief that these high capacity transporters are responsible for modulating chemosensitvity of colorectal cancer. Several immunohistochemistry and reverse transcription polymerase chain reaction (RT-PCR) studies have provided controversial results in regards to the expression levels of these two ABC-transporters in colorectal cancer. Our study was designed to determine the yet uninvestigated functional activity of MDR1 and MRP1 transporters in normal human enterocytes compared to colorectal cancer cells from surgical biopsies. Methods 100 colorectal cancer and 28 adjacent healthy mucosa samples were obtained by intraoperative surgical sampling. Activity of MDR1 and MRP1 of viable epithelial and cancer cells were determined separately with the modified calcein-assay for multidrug resistance activity and sufficient data of 73 cancer and 11 healthy mucosa was analyzed statistically. Results Significantly decreased mean MDR1 activity was found in primary colorectal cancer samples compared to normal mucosa, while mean MRP1 activity showed no significant change. Functional activity was not affected by gender, age, stage or grade and localization of the tumor. Conclusion We found lower MDR activity in cancer cells versus adjacent, apparently, healthy control tissue, thus, contrary to general belief, MDR activity seems not to play a major role in primary drug resistance, but might rather explain preferential/selective activity of Irinotecan and/or Oxaliplatin. Still, this picture might be more complex since chemotherapy by itself might alter MDR activity, and furthermore, today limited data is available about MDR activity of cancer stem cells in colorectal cancers. Virtual slides The virtual slide(s) for this article can be found here: http://www.diagnosticpathology.diagnomx.eu/vs/1675739129145824

In this paper we propose a novel architecture and approach to provide accountability for Web communities that require a high-level of privacy. A two-layered privacy protection architecture is proposed, that supports (i) registration of participants and enforcement of community rules, called internal accountability , and (ii) rule-based interaction with real world organizations, called external accountability . Our security protocols build upon community-based trust and limit the exposure of private data on trusted third parties. The two-layered architecture protects the mappings between real users and their virtual identities, and among the virtual users, while guaranteeing internal and external accountability. We target Web communities that are dynamic and self-organizing, i.e. roles and contributions of participants may change over time. The proposed concepts and protocols are implemented in our SyllabNet project that supports anonymous course evaluations by university students.

The slogan that information is powerhas undergone a slight change. Today, information updatingis in the focus of interest. The largest source of information is the world-wide web. Fast search methods are in need for this enormous source. In this paper a hybrid architecture that combines soft support vector classification and reinforcement learning for value estimation is introduced for the evaluation of a link (a document) and its neighboring links (or documents), called the context of a document. The method is motivated by (i) large differences between such contexts on the web, (ii) the facilitation of goal oriented search using context classifiers, and (iii) attractive fast adaptation properties, that could counteract diversity of web environments. We demonstrate that value estimation-based fast adaptation offers considerable improvement over other known search methods. © 2002 Elsevier Science B.V. All rights reserved.

Cikkemben érveket hozok fel amellett, hogy, hogy a technológiai fejlődés ma nagy lehetőségeket kínálnak az egészségügy és a jólét számára. Nézetem szerint (1) az „okos” eszközök (smart tools) és a különböző viselhető érzékelők, (2) az adatgyűjtés és az adatbányászati módszerek, (3) a három dimenziós (3D-s) képi rögzítési és képi feldolgozási eszközök, (4) a 3D-s, bonyolult fizikai motorral rendelkező, például grafikai modellek, valamint (5) a crowdsourcing-on (outsourcing: külső erőforrások igénybevétele, crowdsourcing: külső emberi erőforrások tömeges igénybevétele) alapuló emberalapú számítások (human-based computing), terén történő nagy és sikeres erőfeszítések hatalmas változásokat indítanak el. Nem állítom, bár tagadni sem tudom azt, hogy a mesterséges intelligencia eszközei néhány év múlva elérik az emberi intelligencia szintjét, mert ez lehetséges. Véleményem szerint, az egészségügy és a jólét területén gyors fejlődés lehetséges az egészségügyi és jóléti szakértők, és a motivált mérnökök közötti aktív együttműködés útján. --- Artificial Intelligence, Health and Wellbeing: prospects for machine learning, crowdsourcing and self-annotation We argue that recent technology developments – e.g. smart tools and wearable sensors of diverse kinds, data collection and data mining methods, 3D visual recording and visual processing methods, 3D models of the environment with robust physics engine – and new applications of human computing and crowdsourcing hold great promises for health and wellbeing. We are neither claiming nor excluding that human intelligence will be reached in some years from now, but make the above claim, which is both weaker and stronger. We believe that fast developments for health and wellbeing are the question of active collaboration between health and wellbeing experts and motivated engineers.

Effects of ursolic acid on the structural and morphological characteristics of dipalmitoyl lecithin(DPPC)-water system was studied by using differential scanning calorimetry (DSC), small- and wide-angle X-ray scattering (SWAXS), freeze-fracture method combined with transmission electron-microscopy (FF-TEM) and infrared spectroscopy (FT-IR). The surface of the uncorrelated lipid system is rippled or grained and a huge number of small, presumably unilamellar vesicles are present if the UA/DPPC molar ratio is 0.1 mol/mol or higher. Besides the destroyed layer packing of regular multilamellar vesicles, non-bilayer (e.g. cubic or hexagonal) local structures are evidenced by SAXS and FF-TEM methods. The ability of UA to induce non-bilayer structures in hydrated DPPC system originates from the actual geometry form of associated lipid and UA molecules as concluded from the FT-IR measurements and theoretical calculations. Beside numerous beneficial e.g. chemopreventive and chemotherapeutic effect of ursolic acid against cancer, their impact to modify the lipid bilayers can be utilized in liposomal formulations.

In many behavioral domains, such as facial expression and gesture, sparse structure is prevalent. This sparsity would be well suited for event detection but for one problem. Features typically are confounded by alignment error in space and time. As a consequence, high-dimensional representations such as SIFT and Gabor features have been favored despite their much greater computational cost and potential loss of information. We propose a Kernel Structured Sparsity (KSS) method that can handle both the temporal alignment problem and the structured sparse reconstruction within a common framework, and it can rely on simple features. We characterize spatio-temporal events as time-series of motion patterns and by utilizing time-series kernels we apply standard structured-sparse coding techniques to tackle this important problem. We evaluated the KSS method using both gesture and facial expression datasets that include spontaneous behavior and differ in degree of difficulty and type of ground truth coding. KSS outperformed both sparse and non-sparse methods that utilize complex image features and their temporal extensions. In the case of early facial event classification KSS had 10% higher accuracy as measured by F1 score over kernel SVM methods1.

Recently, computer simulation aided work has become a standard routine in all engineering fields. Accordingly, simulation plays a fundamental role even in road traffic engineering. A reliable simulator is able to provide effective analysis of a given traffic network if the applied simulation scenarios properly converge to the real-word situation. This requirement can be achieved based on the mixed use of prior real-world traffic measurements and proper simulation settings. The latter one, however, is not straightforward. Accordingly, the paper investigates a potential calibration technique to create realistic simulations. Basically, a tuning method with genetic algorithm is proposed to reproduce true-to-life traffic based on floating car speed data.

A brain model-based alternative to reinforcement learning is presented that integrates artificial neural networks and knowledge-based systems into one unit or agent for goal-oriented problem solving. The agent may possess inherited and learned artificial neural networks and knowledge-based subsystems. The agent has and develops ANN cues to the environment for dimensionality reduction (data compression) to ease the problem of combinatorial explosion. Here, a dynamical concept model is put forward that builds cue models of the phenomena in the world, designs dynamical action sets (concepts), and makes them compete in a spreading-activation neural stage to reach decision. The agent works under closed-loop control. Here we examine a simple robotlike object in a two-dimensional conditionally probabilistic space.

The estimation of relevant information theoretical quantities, such as entropy, mutual information, and various divergences is computationally expensive in high dimensions. However, for this task, one may apply pairwise Euclidean distances of sample points, which suits random projection (RP) based low dimensional embeddings. The Johnson-Lindenstrauss (JL) lemma gives theoretical bound on the dimension of the low dimensional embedding. We adapt the RP technique for the estimation of information theoretical quantities. Intriguingly, we find that embeddings into extremely small dimensions, far below the bounds of the JL lemma, provide satisfactory estimates for the original task. We illustrate this in the Independent Subspace Analysis (ISA) task; we combine RP dimension reduction with a simple ensemble method. We gain considerable speed-up with the potential of real-time parallel estimation of high dimensional information theoretical quantities.

We treat the problem of searching for hidden multi-dimensional independent auto-regressive processes. First, we transform the problem to Independent Subspace Analysis (ISA). Our main contribution concerns ISA. We show that under certain conditions, ISA is equivalent to a combinatorial optimization problem. For the solution of this optimization we apply the cross-entropy method. Numerical simulations indicate that the cross-entropy method can provide considerable improvements over other state-of-the-art methods.

Independent subspace analysis (ISA) that deals with multi-dimensional independent sources, is a generalization of independent component analysis (ICA). However, all known ISA algorithms may become ineffective when the sources possess temporal structure. The innovation process instead of the original mixtures has been proposed to solve ICA problems with temporal dependencies. Here we show that this strategy can be applied to ISA as well. We demonstrate the idea on a mixture of 3D processes and also on a mixture of facial pictures used as two-dimensional deterministic sources. ISA on innovations was able to find the original subspaces, while plain ISA was not.

This paper introduces an unsupervised algorithm that collects senses contained in WordNet to explain words, whose meaning is unknown, but plenty of documents are available that contain the word in that unknown sense. Based on the widely accepted idea that the meaning of a word is characterized by its context, a neural network architecture was designed to reconstruct the meaning of the unknown word. The connections of the network were derived from word co-occurrences and word-sense statistics. The method was tested on 80 TOEFL synonym questions, from which 63 questions were answered correctly. This is comparable to other methods tested on the same questions, but using a larger corpus or richer lexical database. The approach was found robust against details of the architecture.

In this paper we compare our selection based learning algorithm with the reinforcement learning algorithm in Web crawlers. The task of the crawlers is to find new information on the Web. We performed simulations based on data collected from the Web. The collected portion of the Web is typical and exhibits scale-free small world (SFSW) structure. We have found that on this SFSW, the weblog update algorithm performs better than the reinforcement learning algorithm. It finds the new information faster than the reinforcement learning algorithm and has better new information/all submitted documents ratio.

We propose a method to determine unintelligible words based on the textual context of the word determined. As there can be many different possibilities for the word, a robust, large-scale method is needed. The large scale makes the problem sensitive to spurious similarities of contexts: when the contexts of two, different words are similar. To reduce this effect, we induce structured sparsity on the words by formulating the task as a group Lasso problem. We compare this formulation to a k-nearest neighbor and a support vector machine based approach, and find that group Lasso outperforms both by a large margin. We achieve up to 75% of accuracy when determining the word from among 1000 words both on the Brown corpus and on the British National Corpus. Unintelligible words are often the result of errors in Optical Character Recognition (OCR) algorithms. As the proposed method utilizes information independent from information used in OCR, we expect that a combined approach could be very successful, as OCR and the proposed method complement each other.

Background Rheumatoid Arthritis is a chronic disease leading to decreased quality of life with a rather variable response rate to Disease Modifying Anti Rheumatic Drugs. Methotrexate (MTX) is the gold standard therapy in Rheumatoid Arthritis. The Multidrug resistance Related Protein and Multi Drug Resistance protein 1, also called P-glycoprotein-170 transporters can alter the intracellular concentration of different drugs. Methotrexate is an MRP1 substrate and thus the functional activity of MRP1 might have a clinical impact on the efficiency of the Methotrexate-therapy in Rheumatoid Arthritis. Methods We have compared the functional Multidrug Activity Factors (MAF) of the MDR1 and MRP1 transporters of Peripheral Blood Leukocytes of 59 Rheumatoid Arthritis patients with various response rate to MTX-therapy (MTX-responder, MTX-resistant and MTX-intolerant RA-groups) and 47 non-RA controls in six different leukocyte subpopulations (neutrophil leukocytes, monocytes, lymphocytes, CD4+, CD8+ and CD19+ cells). There was a decreased MAF of RA patients compared to non- Rheumatoid Arthritis patients and healthy controls in the leukocyte subpopulations. There was a significant difference between the MAF values of the MTX-responder and MTX intolerant groups. But we have not found significant differences between the MAF values of the MTX-responder and MTX-resistant Rheumatoid Arthritis -groups. Results Our results suggest that MDR1 and MRP1 functional activity does not seem to affect the response rate to MTX-therapy of Rheumatoid Arthritis-patients, but it might be useful in predicting MTX-side effects. We have demonstrated the decreased functional MDR-activity on almost 60 Rheumatoid Arthritis patients, which can be interpreted as a sign of the immune-suppressive effect of the MTX-treatment.