Ranking Task Research Articles

Analysis on the potential of an EA–surrogate modelling tandem for deep learning parametrization: an example for cancer classification from medical images

The paper introduces a novel modality to efficiently tune the convolutional layers of a deep neural network (CNN) and an approach to also rank the importance of the involved hyperparameters. Evolutionary algorithms (EA) offer a flexible solution to this twofold issue, while the expensive simulations of the deep learner with the generated configurations are resolved by surrogate modelling. Three models have been used and evaluated as surrogates: random forests (RF), support vector machines (SVM) and Kriging. Sample convolutional configurations are generated by Latin hypercube sampling and have attached computed accuracy outcomes from real CNN runs. For the hyperparameter estimation task, the fitness of an individual from the EA associated with a surrogate model is subsequently derived from the CNN accuracy estimation on those variable values. With respect to the ranking and variable selection task, RF includes implicit variable selection, the SVM can be straightforwardly supported by a second EA, and Kriging offers a ranking based on the corresponding θ values. The estimated accuracy of the found hyperparameter values is compared with the true validation accuracy, and they are next used for the prediction on the test cases. The ranking of the variables for each of the three surrogate models is compared, and their influence is also revealed by response surface methodology. The experimental testing of the proposed EA–surrogate approaches is conducted on a real-world scenario of histopathological image interpretation in colorectal cancer diagnosis.

Cross-sectional study of students’ knowledge of sizes and distances of astronomical objects

This paper reports on the results from administering a modified version of the Introductory Astronomy Questionnaire (IAQ) to middle school students and preservice science teachers in Norway. Ranking tasks formed a key part of the instrument, and we detail a new method for analyzing ranking task data. One of our main findings was that even after instruction, a significant proportion of students held erroneous views and conceptions regarding sizes, distances, and the nature of basic astronomical entities, such as stars and planets. We argue that the commonalities between some of the issues we identified and those presented in extant studies—from a variety of countries, with samples ranging from junior high school students and undergraduates to primary school teachers—may point to deeper cognitive issues inherent in, and possibly unique to, engaging with astronomy.

A Q-Convexity Vector Descriptor for Image Analysis

Shape representation is a main problem in computer vision, and shape descriptors are widely used for image analysis. In this paper, based on the previous work Balazs, P., Brunetti, S.: A New Shape Descriptor Based on a Q-convexity Measure, Lecture Notes in Computers Science 10502, 20th Discrete Geometry for Computer Imagery (DGCI) (2017) 267–278, we design a new convexity vector descriptor derived by the notion of the so-called generalized salient points matrix. We investigate properties of the vector descriptor, such as scale invariance and its behavior in a ranking task. Moreover, we present results on a binary and a multiclass classification problem using k-nearest neighbor, decision tree, and support vector machine methods. Results of these experiments confirm the good behavior of our proposed descriptor in accuracy, and its performance is comparable and, in some cases, superior to some recently published similar methods.

A Mixed Methods Approach to Understanding Antiretroviral Treatment Preferences: What Do Patients Really Want?

As the number of effective antiretroviral therapy (ART) options increases, there is greater opportunity to involve HIV patients in ART selection. To establish the parameters for a shared ART decision-making process, we sought to identify ART characteristics that are important to patients and understand considerations in ART selection from both patient and provider perspectives. Using a mixed-methods approach, 16 patients and 12 healthcare providers were interviewed, and ranking tasks were conducted with 26 patients to identify ART characteristics that are relevant for shared decision making. Interviews were coded using direct content analysis and complemented by a quantitative analysis of references to specific attributes. Ranking data were analyzed through count analysis. Qualitative analysis of patient interviews identified four major categories of ART characteristics that are pertinent to shared decision making: side effects (14/16 patients), administration characteristics (14/16 patients), dosing (13/16 patients), and long-term effects (12/16 patients). Other considerations included expectations for patient involvement in ART decision making, relationship with provider, and efficacy. The degree of concordance between patients and providers differed across categories. Ranking exercises demonstrated differences in the ways providers and patients prioritize specific side effects and food requirements. Expectations for patient involvement in the selection process also varied greatly among and between patients and providers. We identify specific attributes of ART that are decision-relevant to patients and providers, describe heterogeneity of their relative importance, and note variable perceptions of shared decision making. Individualizing ART will require greater investment in understanding an individual patient's preferences, including her/his desire to participate in shared decision making.

Altered neural responsivity to food cues in relation to food preferences, but not appetite-related hormone concentrations after RYGB-surgery

BackgroundAfter Roux-en-Y gastric bypass (RYGB) surgery, patients report a shift in food preferences away from high-energy foods. ObjectiveWe aimed to elucidate the potential mechanisms underlying this shift in food preferences by assessing changes in neural responses to food pictures and odors before and after RYGB. Additionally, we investigated whether altered neural responsivity was associated with changes in plasma endocannabinoid and ghrelin concentrations. Design19 RYGB patients (4 men; age 41 ± 10 years; BMI 41 ± 1 kg/m2 before; BMI 36 ± 1 kg/m2 after) participated in this study. Before and two months after RYGB surgery, they rated their food preferences using the Macronutrient and Taste Preference Ranking Task and BOLD fMRI responses towards pictures and odors of high-, and low-energy foods and non-food items were measured. Blood samples were taken to determine plasma endocannabinoid and ghrelin concentrations pre- and post-surgery. ResultsPatients demonstrated a shift in food preferences away from high-fat/sweet and towards low-energy/savory food products, which correlated with decreased superior parietal lobule responsivity to high-energy food odor and a reduced difference in precuneus responsivity to high-energy versus low-energy food pictures. In the anteroventral prefrontal cortex (superior frontal gyrus) the difference in deactivation towards high-energy versus non-food odors reduced. The precuneus was less deactivated in response to all cues. Plasma concentrations of anandamide were higher after surgery, while plasma concentrations of other endocannabinoids and ghrelin did not change. Alterations in appetite-related hormone concentrations did not correlate with changes in neural responsivity. ConclusionsRYGB leads to changed responsivity of the frontoparietal control network that orchestrates top-down control to high-energy food compared to low-energy food and non-food cues, rather than in reward related brain regions, in a satiated state. Together with correlations with the shift in food preference from high- to low-energy foods this indicates a possible role in new food preference formation.

An efficient top-k ranking method for service selection based on ε-ADMOPSO algorithm

One of the main concerns in rank aggregation tasks for metasearch service is how to retrieve and aggregate the large-scale candidate search results efficiently. Much work has been done to implement metasearch service engines with different rank aggregation algorithms. However, the performance of these metasearch engines can hardly be improved. In this paper, we transform the top-k ranking task into a multi-objective programming problem when user preferences are considered along with user queries. We build an improved discrete multi-objective programming model to make the aggregate rankings satisfy user queries and user preferences both, and then propose a user preferences-based rank aggregation algorithm accordingly. Based on discrete particle swarm optimization algorithm, we improve the encoding scheme, the initialization methods, the position and velocity definition, the integrating updating process, the turbulence operator, and the external archive updating and leader selection strategy to make sure the candidate results that fit the user’s preferences can be located quickly and accurately in a large-scale discrete solution space. We have our proposed algorithm tested on three different benchmark datasets: a public dataset, the real-world datasets and the synthetic simulation datasets. The experimental results demonstrate the efficacy and convergence efficiency of the proposed algorithm over the baseline rank aggregation methods especially when dealing with large amount of candidate results. And when the set of candidate results is of normal size, the proposed algorithm is proved to perform not worse than the baseline methods.

Bernoulli Embeddings for Graphs

Just as semantic hashing can accelerate information retrieval, binary valued embeddings can significantly reduce latency in the retrieval of graphical data. We introduce a simple but effective model for learning such binary vectors for nodes in a graph. By imagining the embeddings as independent coin flips of varying bias, continuous optimization techniques can be applied to the approximate expected loss. Embeddings optimized in this fashion consistently outperform the quantization of both spectral graph embeddings and various learned real-valued embeddings, on both ranking and pre-ranking tasks for a variety of datasets.

StarSpace: Embed All The Things!

We present StarSpace, a general-purpose neural embedding model that can solve a wide variety of problems: labeling tasks such as text classification,ranking tasks such as information retrieval/web search,collaborative filtering-based or content-based recommendation,embedding of multi-relational graphs, and learning word, sentence or document level embeddings.In each case the model works by embedding those entities comprised of discrete features and comparing them against each other -- learning similarities dependent on the task.Empirical results on a number of tasks show that StarSpace is highly competitive with existing methods, whilst also being generally applicable to new cases where those methods are not.

S-Net: From Answer Extraction to Answer Synthesis for Machine Reading Comprehension

In this paper, we present a novel approach to machine reading comprehension for the MS-MARCO dataset. Unlike the SQuAD dataset that aims to answer a question with exact text spans in a passage, the MS-MARCO dataset defines the task as answering a question from multiple passages and the words in the answer are not necessary in the passages. We therefore develop an extraction-then-synthesis framework to synthesize answers from extraction results. Specifically, the answer extraction model is first employed to predict the most important sub-spans from the passage as evidence, and the answer synthesis model takes the evidence as additional features along with the question and passage to further elaborate the final answers. We build the answer extraction model with state-of-the-art neural networks for single passage reading comprehension, and propose an additional task of passage ranking to help answer extraction in multiple passages. The answer synthesis model is based on the sequence-to-sequence neural networks with extracted evidences as features. Experiments show that our extraction-then-synthesis method outperforms state-of-the-art methods.

Attitudes to variation in spoken Faroese

This paper presents results from a language attitude test in the Faroe Islands, a small country with considerable dialectal variation and no standard spoken language. The test used is a replica of the verbal guise test method that has been used through the SLICE research network, which includes a verbal guise test to elicit reactions to five regional varieties of Faroese and a label ranking task to elicit statements on language variation. There were 203 Faroese 15‐year‐old students tested. The results from the label ranking task indicate a common notion of a national Faroese dialect hierarchy that favors that of the capital. However, the reaction patterns in the verbal guise test reflect no such hierarchy. The Faroese results thus diverge from the evaluation patterns found in a number of comparable European verbal guise tests, yet are similar to the patterns found in one European test locality: Western Norway.

Combinative hypergraph learning in subspace for cross-modal ranking

Recent years have witnessed a surge of interests in cross-modal ranking. To bridge the gap between heterogeneous modalities, many projection based methods have been studied to learn common subspace where the correlation across different modalities can be directly measured. However, these methods generally consider pair-wise relationship merely, while ignoring the high-order relationship. In this paper, a combinative hypergraph learning in subspace for cross-modal ranking (CHLS) is proposed to enhance the performance of cross-modal ranking by capturing high-order relationship. We formulate the cross-modal ranking as a hypergraph learning problem in latent subspace where the high-order relationship among ranking instances can be captured. Furthermore, we propose a combinative hypergraph based on fused similarity information to encode both the intra-similarity in each modality and the inter-similarity across different modalities into the compact subspace representation, which can further enhance the performance of cross-modal ranking. Experiments on three representative cross-modal datasets show the effectiveness of the proposed method for cross-modal ranking. Furthermore, the ranking results achieved by the proposed CHLS can recall 80% of the relevant cross-modal instances at a much earlier stage compared against state-of-the-art methods for both cross-modal ranking tasks, i.e. image query text and text query image.

Terrorism Watch Lists, Suspect Ranking and Decision-Making Biases

ABSTRACTThe large number of names on terrorism watch lists raises the problem of monitoring. Given the existing resource constraints and other logistical considerations, efficient and accurate ranking of individuals in terms of threat posed is of paramount importance. This process, however, may be impacted by reference points, diminishing sensitivity, loss aversion, and other aspects of the human decision-making process that introduce biases. This article explores the relevance of decision-making processes and biases to the specific task of ranking and monitoring individuals whose names have been placed on a terrorism watch list.

Multi‐stage ranking approach for fast person re‐identification

One of the goals of person re-identification systems is to support video-surveillance operators and forensic investigators to find an individual of interest in videos acquired by a network of non-overlapping cameras. This is attained by sorting images of previously observed individuals for decreasing values of their similarity with a given probe individual. Existing appearance descriptors, together with their similarity measures, are mostly aimed at improving ranking quality. The authors address instead the issue of processing time, which is also relevant in practical applications involving interaction with human operators. They show how a trade-off between processing time and ranking quality, for any given descriptor, can be achieved through a multi-stage ranking approach inspired by multi-stage classification approaches, which they adapt to the re-identification ranking task. The authors analytically model the processing time of multi-stage system and discuss the corresponding accuracy, and derive from these results practical design guidelines. They then empirically evaluate their approach on three benchmark data sets and four state-of-the-art descriptors.

Comparison of rapid descriptive sensory methodologies: Free-Choice Profiling, Flash Profile and modified Flash Profile

Rapid sensory methods have been developed as alternatives to traditional sensory descriptive analysis methods. Among them, Free-Choice Profiling (FCP) and Flash Profile (FP) are two that have been known for many years. The objectives of this work were to compare the rating-based FCP and ranking-based FP method; to evaluate the impact of adding adjustments to FP approach; to investigate the influence of the number of assessors on the outcome of modified FP. To achieve these aims, a conventional descriptive analysis (DA), FCP, FP and a modified version of FP were carried out. Red wines made by different grape maturity and ethanol concentration were used for sensory testing. This study showed that DA provided a more detailed and accurate information on products through a quantitative measure of the intensity of sensory attributes than FCP and FP. However, the panel hours for conducting DA were higher than that for rapid methods, and FP was even able to separate the samples to a higher degree than DA. When comparing FCP and FP, this study showed that the ranking-based FP provided a clearer separation of samples than rating-based FCP, but the latter was an easier task for most assessors. When restricting assessors on their use of attributes in FP, the sample space became clearer and the ranking task was simplified. The FP protocol with restricted attribute sets seems to be a promising approach for efficient screening of sensory properties in wine. When increasing the number of assessors from 10 to 20 for conducting the modified FP, the outcome tended to be slightly more stable, however, one should consider the degree of panel training when deciding the optimal number of assessors for conducting FP.

Load Balancing Scheduling Algorithm for Concurrent Workflow

Concurrent workflow scheduling algorithm works in three phases, namely rank computation, tasks selection, and resource selection. In this paper, we introduce a new ranking algorithm that computes the rank of a task, based on its successor rank and its predecessors average communication time, instead of its successors rank. The advantage of this ranking algorithm is that two dependent tasks are assigned to the same machine and as a result the scheduled length is reduced. The task selection phase selects a ready task from each workflow and creates a task pool. The resource selection phase initially assigns tasks using min-min heuristic, after initial assignment, tasks are moved from the highly loaded machines to the lightly loaded machines. Our resource selection algorithm increases the load balance among the resources due to tasks assignment heuristic and reassignment of tasks from the highly loaded machines. The simulation results show that our proposed scheduling algorithm performs better over existing approaches in terms of load balance, makespan and turnaround time.

Content-Oriented User Modeling for Personalized Response Ranking in Chatbots

Automatic chatbots (also known as chat-agents) have attracted much attention from both researching and industrial fields. Generally, the semantic relevance between users’ queries and the corresponding responses is considered as the essential element for conversation modeling in both generation and ranking based chat systems. By contrast, it is a nontrivial task to adopt the users’ information, such as preference, social role, etc., into conversational models reasonably, while users’ profiles play a significant role in the procedure of conversations by providing the implicit contexts. This paper aims to address the personalized response ranking task by incorporating user profiles into the conversation model. In our approach, users’ personalized representations are latently learned from the contents posted by them via a two-branch neural network. After that, a deep neural network architecture is further presented to learn the fusion representation of posts, responses, and personal information. In this way, the proposed model could understand conversations from the users’ perspective; hence, the more appropriate responses are selected for a specified person. The experimental results on two datasets from social network services demonstrate that our approach is hopeful to represent users’ personal information implicitly based on user generated contents, and it is promising to perform as an important component in chatbots to select the personalized responses for each user.

Low reported taste function is associated with low preference for high protein products in advanced oesophagogastric cancer patients undergoing palliative chemotherapy

Cancer patients undergoing palliative chemotherapy can experience a variety of chemosensory and food preference changes which may impact their nutritional status and quality of life. However, evidence of these changes in oesophagogastric cancer (OGC) patients is currently mostly qualitative and not supported by quantitative data. The aim of this study was to assess how self-reported and objective taste and smell function and food preferences change over time during chemotherapy in OGC patients. This observational study included 15 advanced OGC patients planned for first line treatment with capecitabine and oxaliplatin. Participants completed two test sessions scheduled before start of cytotoxic treatment and after two cycles. Self-reported and objective taste and smell function and the macronutrient and taste preference ranking task were conducted at each test session. Self-reported taste and smell did not change upon chemotherapy. Objective taste function decreased during chemotherapy, although this was not statistically significant (p=0.06), objective smell function did not change. Before and during chemotherapy, high protein foods were preferred over high carbohydrate and over low energy products, but food preferences did not change over time. A lower self-reported taste function correlated with a lower preference for high-protein products (ρ=0.526, p=0.003). This study suggests that objective taste function decreases during chemotherapy in OGC patients, but not smell function. A low reported taste function was related to a lower preference for high-protein products.

Deadline-Constrained Cost Optimization Approaches for Workflow Scheduling in Clouds

Nowadays it is becoming more and more attractive to execute workflow applications in the cloud because it enables workflow applications to use computing resources on demand. Meanwhile, it also challenges traditional workflow scheduling algorithms that only concentrate on optimizing the execution time. This paper investigates how to minimize execution cost of a workflow in clouds under a deadline constraint and proposes a metaheuristic algorithm L-ACO as well as a simple heuristic ProLiS. ProLiS distributes the deadline to each task, proportionally to a novel definition of probabilistic upward rank, and follows a two-step list scheduling methodology: rank tasks and sequentially allocates each task a service which meets the sub-deadline and minimizes the cost. L-ACO employs ant colony optimization to carry out deadline-constrained cost optimization: the ant constructs an ordered task list according to the pheromone trail and probabilistic upward rank, and uses the same deadline distribution and service selection methods as ProLiS to build solutions. Moreover, the deadline is relaxed to guide the search of L-ACO towards constrained optimization. Experimental results show that compared with traditional algorithms, the performance of ProLiS is very competitive and L-ACO performs the best in terms of execution costs and success ratios of meeting deadlines.

Effect of Task Through Test Rankings Based Learning Model REACT to the Achievement of Physics

The aim of this study is to find out the effect of test-based on the ranking task through the learning model of REACT (relating, experiencing, applying, cooperating, and transferring) toward physics learning achievement. The research is held in 105 Senior High School in Jakarta on the even semester of 2015-2016. The population in this research is around 32 students roommates are all investigated. The sampling technique that use is purposing sampling technique. This research use quantitative research. After the data is distributed normally and the homogeneity based on the normality test and homogeneity, the data that is gained the lcount < ltable (0.11 < 0.156). The hypothesis test is done by using the t-test and the result tcount > ttable (8.575 > 1.689) (on the 0.05 significant extent) or 8.575 > 2.456 (on the 0.01 significant extent) which means H0 is rejected, H1 is therefore accepted. All in all this research could bring the effect of test-based on the ranking task learning model through the REACT toward physics learning achievement.

Translucence perception is not dependent on cortical areas critical for processing colour or texture

Translucence is an important property of natural materials, and human observers are adept at perceiving changes in translucence. Perceptions of different material properties appear to arise from different cortical regions, and it is therefore plausible that the perception of translucence is dependent on specialised regions, separate from those important for colour and texture processing. To test for anatomical independence between areas necessary for colour, texture and translucence perception we assessed translucency perception in a cortically colour blind observer, who performs at chance on tasks of colour and texture discrimination. Firstly, in order to establish that MS has shown no significant recovery, we assessed his colour perception performance on the Farnsworth-Munsell 100 Hue Test. Secondly, we tested him with two translucence ranking tasks. In one task, stimuli were images of glasses of tea varying in tea strength. In the other, stimuli were glasses of tea varying only in milkiness. MS was able to systematically rank both strength and milkiness, although less consistently than controls, and for tea strength his rankings were in the opposite order. An additional group of controls tested with greyscale versions of the images succeeded at the tasks, albeit slightly less consistently on the milkiness task, showing that the performance of normal observers cannot be transformed into the performance of MS simply by removing colour information from the stimuli. The systematic performance of MS suggests that some aspects of translucence perception do not depend on regions critical for colour and texture processing.