Technion – Israel Institute of Technology

About: Technion – Israel Institute of Technology is a education organization based out in Haifa, Israel. It is known for research contribution in the topics: Population & Upper and lower bounds. The organization has 31714 authors who have published 79377 publications receiving 2603976 citations. The organization is also known as: Technion Israel Institute of Technology & Ṭekhniyon, Makhon ṭekhnologi le-Yiśraʼel.

A Local-Ratio Theorem for Approximating the Weighted Vertex Cover Problem

Abstract: A local-ratio theorem for approximating the weighted vertex cover problem is presented. It consists of reducing the weights of vertices in certain subgraphs and has the effect of local-approximation. Putting together the Nemhauser-Trotter local optimization algorithm and the local-ratio theorem yields several new approximation techniques which improve known results from time complexity, simplicity and performance-ratio point of view. The main approximation algorithm guarantees a ratio of where K is the smallest integer s.t. † This is an improvement over the currently known ratios, especially for a “practical” number of vertices (e.g. for graphs which have less than 2400, 60000, 10 12 vertices the ratio is bounded by 1.75, 1.8, 1.9 respectively).

Shrinkage Algorithms for MMSE Covariance Estimation

Abstract: We address covariance estimation in the sense of minimum mean-squared error (MMSE) when the samples are Gaussian distributed. Specifically, we consider shrinkage methods which are suitable for high dimensional problems with a small number of samples (large p small n). First, we improve on the Ledoit-Wolf (LW) method by conditioning on a sufficient statistic. By the Rao-Blackwell theorem, this yields a new estimator called RBLW, whose mean-squared error dominates that of LW for Gaussian variables. Second, to further reduce the estimation error, we propose an iterative approach which approximates the clairvoyant shrinkage estimator. Convergence of this iterative method is established and a closed form expression for the limit is determined, which is referred to as the oracle approximating shrinkage (OAS) estimator. Both RBLW and OAS estimators have simple expressions and are easily implemented. Although the two methods are developed from different perspectives, their structure is identical up to specified constants. The RBLW estimator provably dominates the LW method for Gaussian samples. Numerical simulations demonstrate that the OAS approach can perform even better than RBLW, especially when n is much less than p . We also demonstrate the performance of these techniques in the context of adaptive beamforming.

Development, learning and memory in large random networks of cortical neurons: lessons beyond anatomy.

Abstract: 1. Introduction 631.1 Outline 631.2 Universals versus realizations in the study of learning and memory 642. Large random cortical networks developing ex vivo 652.1 Preparation 652.2 Measuring electrical activity 673. Spontaneous development 693.1 Activity 693.2 Connectivity 704. Consequences of spontaneous activity: pharmacological manipulations 724.1 Structural consequences 724.2 Functional consequences 735. Effects of stimulation 745.1 Response to focal stimulation 745.2 Stimulation-induced changes in connectivity 746. Embedding functionality in real neural networks 776.1 Facing the physiological definition of ‘reward’: two classes of theories 786.2 Closing the loop 797. Concluding remarks 848. Acknowledgments 859. References 85The phenomena of learning and memory are inherent to neural systems that differ from each other markedly. The differences, at the molecular, cellular and anatomical levels, reflect the wealth of possible instantiations of two neural learning and memory universals: (i) an extensive functional connectivity that enables a large repertoire of possible responses to stimuli; and (ii) sensitivity of the functional connectivity to activity, allowing for selection of adaptive responses. These universals can now be fully realized in ex-vivo developing neuronal networks due to advances in multi-electrode recording techniques and desktop computing. Applied to the study of ex-vivo networks of neurons, these approaches provide a unique view into learning and memory in networks, over a wide range of spatio-temporal scales. In this review, we summarize experimental data obtained from large random developing ex-vivo cortical networks. We describe how these networks are prepared, their structure, stages of functional development, and the forms of spontaneous activity they exhibit (Sections 2–4). In Section 5 we describe studies that seek to characterize the rules of activity-dependent changes in neural ensembles and their relation to monosynaptic rules. In Section 6, we demonstrate that it is possible to embed functionality into ex-vivo networks, that is, to teach them to perform desired firing patterns in both time and space. This requires ‘closing a loop’ between the network and the environment. Section 7 emphasizes the potential of ex-vivo developing cortical networks in the study of neural learning and memory universals. This may be achieved by combining closed loop experiments and ensemble-defined rules of activity-dependent change.

Serpent: A Proposal for the Advanced Encryption Standard

Abstract: We propose a new block cipher as a candidate for the Advanced Encryption Standard. Its design is highly conservative, yet still allows a very efficient implementation. It uses S-boxes similar to those of DES in a new structure that simultaneously allows a more rapid avalanche, a more efficient bitslice implementation, and an easy analysis that enables us to demonstrate its security against all known types of attack. With a 128-bit block size and a 256-bit key, it is as fast as DES on the market leading Intel Pentium/MMX platforms (and at least as fast on many others); yet we believe it to be more secure than three-key triple-DES.