Search

The unbounded knapsack problem with bounded weights is a variant of the well-studied variant of the traditional binary knapsack problem; key changes being the relaxation of the binary constraint and allowing the unit weights of each item to fall within a range. In this paper, we formulate a variant of this problem, which we call the strict unbounded knapsack problem with bounded weights, by replacing the inequality constraint on the total weight with an equality. We show that this problem can be decomposed into a two-stage, pre-emptive goal programming problem, with the first stage being a 2-dimensional knapsack problem and the second being either a linear feasibility program (per canonical formulation) or simply a linearly-constrained program in the general case. This reformulation is shown to be equivalent to the original formulation but allows the use of well-studied, efficient algorithms for multidimensional knapsack problems. In addition, it separates the modeling effort around what to put in the knapsack from considerations around what unit weight one should assign to each item type, providing substantially more flexibility to the modeler without adding complexity to the choice of knapsack configuration. Finally, we show that for the feasibility version of the second stage, one can immediately get a feasible solution to the first stage solution., Comment: 8 pages

We introduce a deep recursive octree network for the compression of 3D voxel data. Our network compresses a voxel grid of any size down to a very small latent space in an autoencoder-like network. We show results for compressing 32, 64 and 128 grids down to just 80 floats in the latent space. We demonstrate the effectiveness and efficiency of our proposed method on several publicly available datasets with three experiments: 3D shape classification, 3D shape reconstruction, and shape generation. Experimental results show that our algorithm maintains accuracy while consuming less memory with shorter training times compared to existing methods, especially in 3D reconstruction tasks.

Any generic deep machine learning algorithm is essentially a function fitting exercise, where the network tunes its weights and parameters to learn discriminatory features by minimizing some cost function. Though the network tries to learn the optimal feature space, it seldom tries to learn an optimal distance metric in the cost function, and hence misses out on an additional layer of abstraction. We present a simple effective way of achieving this by learning a generic Mahalanabis distance in a collaborative loss function in an end-to-end fashion with any standard convolutional network as the feature learner. The proposed method DML-CRC gives state-of-the-art performance on benchmark fine-grained classification datasets CUB Birds, Oxford Flowers and Oxford-IIIT Pets using the VGG-19 deep network. The method is network agnostic and can be used for any similar classification tasks., Comment: arXiv admin note: text overlap with arXiv:1903.09123

We present a conditional probabilistic framework for collaborative representation of image patches. It incorporates background compensation and outlier patch suppression into the main formulation itself, thus doing away with the need for pre-processing steps to handle the same. A closed form non-iterative solution of the cost function is derived. The proposed method (PProCRC) outperforms earlier CRC formulations: patch based (PCRC, GP-CRC) as well as the state-of-the-art probabilistic (ProCRC and EProCRC) on three fine-grained species recognition datasets (Oxford Flowers, Oxford-IIIT Pets and CUB Birds) using two CNN backbones (Vgg-19 and ResNet-50).

We present an end-to-end deep network for fine-grained visual categorization called Collaborative Convolutional Network (CoCoNet). The network uses a collaborative layer after the convolutional layers to represent an image as an optimal weighted collaboration of features learned from training samples as a whole rather than one at a time. This gives CoCoNet more power to encode the fine-grained nature of the data with limited samples. We perform a detailed study of the performance with 1-stage and 2-stage transfer learning. The ablation study shows that the proposed method outperforms its constituent parts consistently. CoCoNet also outperforms few state-of-the-art competing methods. Experiments have been performed on the fine-grained bird species classification problem as a representative example, but the method may be applied to other similar tasks. We also introduce a new public dataset for fine-grained species recognition, that of Indian endemic birds and have reported initial results on it.

The potential lack of fairness in the outputs of machine learning algorithms has recently gained attention both within the research community as well as in society more broadly. Surprisingly, there is no prior work developing tree-induction algorithms for building fair decision trees or fair random forests. These methods have widespread popularity as they are one of the few to be simultaneously interpretable, non-linear, and easy-to-use. In this paper we develop, to our knowledge, the first technique for the induction of fair decision trees. We show that our "Fair Forest" retains the benefits of the tree-based approach, while providing both greater accuracy and fairness than other alternatives, for both "group fairness" and "individual fairness.'" We also introduce new measures for fairness which are able to handle multinomial and continues attributes as well as regression problems, as opposed to binary attributes and labels only. Finally, we demonstrate a new, more robust evaluation procedure for algorithms that considers the dataset in its entirety rather than only a specific protected attribute., Comment: To appear in the first AAAI / ACM conference on Artificial Intelligence, Ethics, and Society (AIES) 2018

This letter introduces the LOOP binary descriptor (local optimal oriented pattern) that encodes rotation invariance into the main formulation itself. This makes any post processing stage for rotation invariance redundant and improves on both accuracy and time complexity. We consider fine-grained lepidoptera (moth/butterfly) species recognition as the representative problem since it involves repetition of localized patterns and textures that may be exploited for discrimination. We evaluate the performance of LOOP against its predecessors as well as few other popular descriptors. Besides experiments on standard benchmarks, we also introduce a new small image dataset on NZ Lepidoptera. Loop performs as well or better on all datasets evaluated compared to previous binary descriptors. The new dataset and demo code of the proposed method are to be made available through the lead author's academic webpage and GitHub.

Binary codes can be used to speed up nearest neighbor search tasks in large scale data sets as they are efficient for both storage and retrieval. In this paper, we propose a robust auto-encoder model that preserves the geometric relationships of high-dimensional data sets in Hamming space. This is done by considering a noise-removing function in a region surrounding the manifold where the training data points lie. This function is defined with the property that it projects the data points near the manifold into the manifold wisely, and we approximate this function by its first order approximation. Experimental results show that the proposed method achieves better than state-of-the-art results on three large scale high dimensional data sets., Comment: Submitting to journal (TPAMI). 17 pages, 11 figures. The Matlab codes for AutoJacoBin and NOKMeans are available: https://bitbucket.org/fxpfxp/autojacobin https://bitbucket.org/fxpfxp/nokmeans The SIFT10M dataset is available at: http://archive.ics.uci.edu/ml/datasets/SIFT10M

A technology professional development initiative was launched in a school district with the goal of extending technology use in the classroom. For teachers to teach expertly, the aim was for them to "be the technology" by modeling technology use in the classroom, applying technology across the curriculum, applying technology to problem-solving and decision-making activities, and applying technology to facilitate collaboration and cooperation among learners. To implement this technology initiative, a set of technology standards and indicators was established to describe best practices for expert teaching and student learning using technology. It was that hypothesized teachers would experience several stages as they developed into expert technology integrators and that this technology initiative could be evaluated by tracking their progress through these stages. This study analyzes the developmental model for technology integration based on the stages, standards, and indicators of the technology professional development initiative and validates its use as a tool for use in the evaluation of technology integration. (Contains 12 references, 3 tables, and 2 figures.) (Author)

What is not fully understood about educational technology is how to transform computer technology into a powerful pedagogical tool. One way to address instructional issues associated with integrating computer technology in classrooms is to analyze educational best practices associated with technology integration in classrooms. Technology standards for teachers were established, and a tool to examine the quality of computer technology integration was developed and validated for a school district undertaking a district-wide technology professional development initiative. The technology standards on which the professional development model was established were formulated by synthesizing national, state, and local technology standards and then identifying educational best practices that supported these standards within the local context. Then a pedagogy was reinforced that enhanced and improved teaching and learning using technology tools and resources through performance assessment and financial incentives. Pattern analysis procedures demonstrated that the measure was appropriate for determining the extent of technology integration and for identifying appropriate technology training themes. (Contains 13 references.) (Author/AEF)

As the use of the Web is perceived to be an effective tool for dissemination of research findings for the provision of asynchronous instruction, the issue of accessibility of Web page information will become more and more relevant. The World Wide Web consortium (W3C) has recognized a disparity in accessibility to the Web between persons with and without disabilities and has responded by developing a set of Web page design standards that specifically address the issue of Web accessibility. The W3C has embraced the issue of accessibility through its Web Accessibility Initiative (WAI). The WAI has developed a set of guidelines, the Web Content Accessibility Guidelines, for use by Web page authors to incorporate accessibility features into the design of Web pages. The purpose of this paper is to describe Web page design principles and strategies based on the WAI's guidelines that will facilitate the development of Web pages that open the gates of the virtual kingdom to persons with disabilities. Web page authoring tips are provided for using the following in a Web page: tables, frames, multimedia, color and backgrounds, and hyperlinks, as well as using alternative Web pages. (AEF)

Schools cannot expect to obtain gains in learner achievement from computer technology if it is not properly implemented. The concerns of teachers in integrating computer technology in the classroom may impact how computer technology is ultimately implemented. This study examined the concerns of elementary school teachers implementing computer-delivered instruction. Data were collected by administering the Stages of Concern Questionnaire (SoCQ) to teachers at four elementary schools in an urban school district. Responses were analyzed for each of the seven stages of concern (i.e., how users perceive an innovation from the time they first become aware of it until they gain mastery of the innovation). The findings of the study support the position that teachers' concerns and perceptions of an integrated learning system (ILS) influence the way in which they implement an ILS. For schools to experience significant change or reform due to the integration of computer technology, the instructional concerns and practices of teachers must be carefully considered. Furthermore, the integration of computer technology must be approved, accepted, and implemented by teachers for computer technology to ultimately impact student learning while reducing instructional costs. (Author/AEF)

This paper describes the development, validation, and research application of the Computer-Delivered Instruction Configuration Matrix (CDICM), an instrument for evaluating the implementation of Integrated Learning Systems (ILS). The CDICM consists of a 15-item checklist, describing the major components of implementation of ILS technology, to be completed by an evaluator based on responses supplied by teachers during an interview. The CDICM was used to examine the operational patterns of teachers (n=30) whose students interacted with an ILS using the "Successmaker" courseware in four elementary schools in a metropolitan school district. Data regarding the operational patterns of teachers using ILS technology were collected from teacher interviews and analyzed. The best ILS implementation practices included integration with classroom instruction, training in the use of an ILS, and the use of motivational strategies. Implications for educational change and reform are discussed. Tables and figures present data on discriminant functions for CDICM components, interview data grouped by implementation pattern, significant CDICM components, and influence variables on degree of ILS implementation. The CDICM and interview guide are included. (Contains 18 references.) (AEF)

This paper examines a research paradigm that is particularly suited to experimentation-related computer-based instruction and integrated learning systems. The main assumption of the model is that one of the most powerful capabilities of computer-based instruction, and specifically of integrated learning systems, is the capacity to adapt instruction to the individual differences that exist among learners. The model is applied to an instructional treatment through a popular integrated learning system from Computer Curriculum Corporation (CCC). The problem of adaptive instruction is to diagnose, in any group of learners, where each is located in relation to individual learning characteristics and the knowledge and performances required at a given point in instruction. The research paradigm to be applied uses a microanalysis technique, based on a regression model proposed by previous researchers, that operationalized variables for each individual. Three figures illustrate the model and discussion. (Contains 49 references.) (SLD)

Discusses theoretical and clinical developments that have accompanied family therapy's entry into the postmodern era. Clinical trends, including use of reflecting teams, self-of-the-therapist issues, increased therapist self-disclosure, and postmodern supervision are examined. Feminist critiques, health-care reform, and increasing collaboration with other mental health professionals are discussed. (JBJ)