Skip to main content

Amihood Amir

Bar-Ilan University, Computer Science, Faculty Member

Followers

9

Following

5

Co-authors

5

Public Views

Professor Amihood Amir received his Ph.D. at Bar-Ilan University in 1983. He was a post-doc at MIT, an Assistant Professor at Tufts University and then at the University of Maryland College Park, and an Associate Professor at Georgia Tech. He has been at Bar-Ilan since 1994. Today he is a Professor at the Department of Computer Science at Bar-Ilan University and Research Professor at the Johns Hopkins University. Professor Amir was head of the Responsa Project, chairman of the Department of Computer Science and dean of the College of Exact Sciences at Bar-Ilan University. He also served as member of the directorate of the Faculty Union at Bar Ilan University and member of the Bar Ilan Senate. He is a member of the Board of Trustees and Board of Directors of Bar Ilan University, and member of the Academic Council of Netanya Academic College. Professor Amir is one of the international leaders in Pattern Matching, an area in which he wrote over 150 papers and book chapters. He is one of the members of the Steering Committee of CPM – the leading international conference in Pattern Matching. He serves on the program committees and editorial boards of leading conferences and journals in Computer Science, and received research grants from, among others, NSF, ISF, BSF, MOS, Magnet, the Royal Society and the EU. He is also recipient of binational grants with South Korea, Finland, France, Italy and the UK. Professor Amir served on panels of the NSF, BSF, and Ministry of Science, and was member and chairman of various accreditation committees of MALAG. He was also chairman of the Computer Science professional committee in the Israel Ministry of Education. Of the 25 Ph.D. students he had supervised in the past, most are now faculty in excellent universities and colleges in Israel and abroad, and his numerous M.Sc. and undergraduate students are ubiquitous in the IT industry. Professor Amir received a best teacher award at the University of Maryland, Research awards at both the University of Maryland and Georgia Tech and a "most inspiring faculty member" award bestowed by the students at the College of Computing at Georgia Tech. In the past, Professor Amir has been project leader in various Ministry of Defense projects and consulted in the real-time systems industry.

less

Robert Van de Geijn

The University of Texas at Austin

Sabanci University

enrique s. quintana-orti

UPV

Imperial College London

Tirthankar Bhattacharyya

Sergio Barrachina

Universitat Jaume I de Castelló

IBM Research

Maria Lucia Sampoli

Interests

Uploads

Papers by Amihood Amir

Asynchronous Pattern Matching

Lecture Notes in Computer Science, 2006

Cheatable, P-terse, and P-superterse Sets

Analyzing high-dimensional data by subspace validity

Third IEEE International Conference on Data Mining

We are proposing a novel method that makes it possible to analyze high dimensional data with arbi... more We are proposing a novel method that makes it possible to analyze high dimensional data with arbitrary shaped projected clusters and high noise levels. At the core of our method lies the idea of subspace validity. We map the data in a way that allows us to test the quality of subspaces using statistical tests. Experimental results, both on synthetic and real data sets, demonstrate the potential of our method.

Parameterized matching on non-linear structures

Information Processing Letters, 2009

The classical pattern matching paradigm is that of seeking occurrences of one string in another, ... more The classical pattern matching paradigm is that of seeking occurrences of one string in another, where both strings are drawn from an alphabet set Σ. In the parameterized pattern matching model, a consistent renaming of symbols from Σ is allowed in a match. The parameterized matching paradigm has proven useful in problems in software engineering, computer vision, and other applications. In classical pattern matching, both the text and pattern are strings. Applications such as searching in xml or searching in hypertext require searching strings in non-linear structures such as trees or graphs. There has been work in the literature on exact and approximate parameterized matching, as well as work on exact and approximate string matching on non-linear structures. In this paper we explore parameterized matching in non-linear structures. We prove that exact parameterized matching on trees can be computed in linear time for alphabets in an O(n)-size integer range, and in time O(n log m) in general, where n is the tree size and m the pattern length. These bounds are optimal in the comparison model. We also show that exact parameterized matching on directed acyclic graphs (DAGs) is N Pcomplete.

Polynomial terse sets

by Amihood Amir and William Gasarch

Information and Computation, 1988

Some connections between bounded query classes and non-uniform complexity

Information and Computation, 2003

Let A(x) be the characteristic function of A. Consider the function F A k (x 1 ,. .. , x k) = A(x... more Let A(x) be the characteristic function of A. Consider the function F A k (x 1 ,. .. , x k) = A(x 1) • • • A(x k). We show that if F A k can be computed with fewer than k queries to some set X then A ∈ P/poly. A generalization of this result has applications to bounded query classes, circuits, and enumerability. In particular we obtain the following. (1) Assuming Σ p 3 = Π p 3 , there are functions computable using f (n) + 1 queries to SAT that are not computable using f (n) queries to SAT, for f (n) = O(log n). (2) If F A k , restricted to length n inputs, can be computed by an unbounded fanin oracle circuit of size s(n) and depth d(n), with k − 1 queries to some set X, then A can be computed with an unbounded fanin (non-oracle) circuit of size n O(k) s(n) and depth d(n) + O(1). (3) Assuming that PH = Σ p 4 ∩ Π p 4 , and < 1, #SAT is not 2 n-enumerable.

Efficient two-dimensional compressed matching

Data Compression Conference, 1992. …, 1992

The problem of two-dimensional run-length compressed matching is defined as follows: Input: Two-d... more

Two-dimensional pattern matching with rotations

Combinatorial Pattern …, 2003

The most efficient currently known algorithms for two dimensional pattern matching with rotations... more The most efficient currently known algorithms for two dimensional pattern matching with rotations have a worst case time complexity of O(n 2 m 3), where the size of the text is n × n and the size of the pattern is m × m. In this paper we present a new algorithm for the problem whose running time is O(n 2 m 2).

Contributions of theoretical computer science

Advanced Query Mechanisms for Biological Databases, Agarwal, Pankaj, Allison, L., Aloy, Patrick,

aaai.org

Advanced Query Mechanisms for Biological Databases,  Agarwal, Pankaj,  Allison, L.,  Aloy, Pa... more Advanced Query Mechanisms for Biological Databases,  Agarwal, Pankaj,  Allison, L.,  Aloy, Patrick,  Altman, Russ B.,  Amir, Amihood,  Ananko, EA,  Atchley, William R.,  Atteson, Kevin,  Automated Clustering and Assembly of Large EST Collections,  Aviles, Francesc X.,  ... Babenko, VN,  Bafna, Vineet,  Baker, Patricia G.,  Baldi, Pierre,  Bayesian Protein Family Classifier,  Bechhofer, Sean,  BioSim—A New Qualitative Simulation Environment for Molecular Biology,  Brass, Andy,  Brunak, ...

Incremental Algorithms for Association Generation

Pacific-Asia Conference on Knowledge Discovery and Data Mining, 1997

Computing Similarity of Run-Length Encoded Strings with Affine Gap Penalty

String Processing and Information Retrieval, 2005

The problem of computing similarity of two run-length encoded strings has been studied for variou... more The problem of computing similarity of two run-length encoded strings has been studied for various scoring metrics. Many algorithms have been developed for the longest common subsequence metric and some algorithms for the Levenshtein distance metric and the weighted edit distance metric. In this paper we consider similarity based on the affine gap penalty metric which is a more general

Fast Parallel and Serial Multidimensional Approximate Array Matching

Sequences, 1990

Improved Deterministic Length Reduction

Computing Research Repository, 2008

This paper presents a new technique for deterministic length reduction. This technique improves t... more This paper presents a new technique for deterministic length reduction. This technique improves the running time of the algorithm presented in (?) for performing fast convolution in sparse data. While the regular fast convolution of vectors V1, V2 whose sizes are N1, N2 respectively, takes O(N1 log N2) using FFT, using the new technique for length reduction, the algorithm proposed

Alphabet independent two dimensional matching

Proceedings of the twenty-fourth annual ACM symposium on Theory of computing - STOC '92, 1992

There are many solutions to the string matching pmZllem whkh are strictly linear in the input siz... more

Optimal two-dimensional compressed matching

Lecture Notes in Computer Science, 1994

Recent proliferation of digitized data and the unprecedented growth in the volume of stored and t... more Recent proliferation of digitized data and the unprecedented growth in the volume of stored and transmitted data motivated the definition of the compressed matching paradigm. This is the problem of efficiently finding a pattern P in a compressed text T without the need to decompress. We present the first optimal two-dimensional compressed matching algorithm. The compression under consideration is the two dimensional run-length compression, used by FAX transmission. We achieve optimal time by proving new properties of two-dimensional periodicity. This enables performing duels in which no witness is required. At the heart of the dueling idea lies the concept that two overlapping occurrences of a pattern in a text can use the content of a predetermined text position or witness in the overlap to eliminate one of them. Finding witnesses is a costly operation in a compressed text, thus the importance of witness-free dueling.

Approximate Matching in Weighted Sequences

Lecture Notes in Computer Science, 2006

Weighted sequences have been recently introduced as a tool to handle a set of sequences that are ... more Weighted sequences have been recently introduced as a tool to handle a set of sequences that are not identical but have many local similarities. The weighted sequence is a “statistical image” of this set, where the probability of every symbol’s occurrence at every text location is given. We address the problem of approximately matching a pattern in such a weighted

Deterministic Length Reduction: Fast Convolution in Sparse Data and Applications

Lecture Notes in Computer Science

In this paper a deterministic algorithm for the length reduction problem is presented. This algor... more In this paper a deterministic algorithm for the length reduction problem is presented. This algorithm enables a new tool for performing fast convolution in sparse data. While the regular fast convolution of vectors V 1 , V 2 whose sizes are N 1 , N 2 respectively, takes O(N 1 log N 2) using FFT, the proposed algorithm performs the convolution in O(n 1 log 3 n 1), where n 1 is the number of non-zero values in V 1. This algorithm assumes that V 1 is given in advance, and the V 2 is given in running time. This running time is achieved using a preprocessing phase on V 1 , which takes O(n 2 1) if N 1 is polynomial in n 1 , and O(n 4 1) if N 1 is exponential in n 1 (which is rarely the case in practical applications). This tool is used to obtain faster results for several well known problems, such as the d-Dimensional Point Set Matching and Searching in Music Archives.

Blocked Pattern Matching Problem and Its Applications in Proteomics

Lecture Notes in Computer Science, 2011

Matching a mass spectrum against a text (a key computational task in proteomics) is slow since th... more Matching a mass spectrum against a text (a key computational task in proteomics) is slow since the existing text indexing algorithms (with search time independent of the text size) are not applicable in the domain of mass spectrometry. As a result, many important applications (e.g., searches for mutated peptides) are prohibitively time-consuming and even the standard search for non-mutated peptides

Efficient randomized dictionary matching algorithms

Combinatorial Pattern Matching, 1992

Abstract The standard string matching problem involves finding all occurrences of a single patter... more Abstract The standard string matching problem involves finding all occurrences of a single pattern in a single text. While this approach works well in many application areas, there are some domains in which it is more appropriate to deal with dictionaries of patterns. A dictionary is a set of patterns; the goal of dictionary matching is to find all dictionary patterns in a given text, simultaneously.

Asynchronous Pattern Matching

Lecture Notes in Computer Science, 2006

Cheatable, P-terse, and P-superterse Sets

Analyzing high-dimensional data by subspace validity

Third IEEE International Conference on Data Mining

We are proposing a novel method that makes it possible to analyze high dimensional data with arbi... more We are proposing a novel method that makes it possible to analyze high dimensional data with arbitrary shaped projected clusters and high noise levels. At the core of our method lies the idea of subspace validity. We map the data in a way that allows us to test the quality of subspaces using statistical tests. Experimental results, both on synthetic and real data sets, demonstrate the potential of our method.

Parameterized matching on non-linear structures

Information Processing Letters, 2009

The classical pattern matching paradigm is that of seeking occurrences of one string in another, ... more The classical pattern matching paradigm is that of seeking occurrences of one string in another, where both strings are drawn from an alphabet set Σ. In the parameterized pattern matching model, a consistent renaming of symbols from Σ is allowed in a match. The parameterized matching paradigm has proven useful in problems in software engineering, computer vision, and other applications. In classical pattern matching, both the text and pattern are strings. Applications such as searching in xml or searching in hypertext require searching strings in non-linear structures such as trees or graphs. There has been work in the literature on exact and approximate parameterized matching, as well as work on exact and approximate string matching on non-linear structures. In this paper we explore parameterized matching in non-linear structures. We prove that exact parameterized matching on trees can be computed in linear time for alphabets in an O(n)-size integer range, and in time O(n log m) in general, where n is the tree size and m the pattern length. These bounds are optimal in the comparison model. We also show that exact parameterized matching on directed acyclic graphs (DAGs) is N Pcomplete.

Polynomial terse sets

by Amihood Amir and William Gasarch

Information and Computation, 1988

Some connections between bounded query classes and non-uniform complexity

Information and Computation, 2003

Let A(x) be the characteristic function of A. Consider the function F A k (x 1 ,. .. , x k) = A(x... more Let A(x) be the characteristic function of A. Consider the function F A k (x 1 ,. .. , x k) = A(x 1) • • • A(x k). We show that if F A k can be computed with fewer than k queries to some set X then A ∈ P/poly. A generalization of this result has applications to bounded query classes, circuits, and enumerability. In particular we obtain the following. (1) Assuming Σ p 3 = Π p 3 , there are functions computable using f (n) + 1 queries to SAT that are not computable using f (n) queries to SAT, for f (n) = O(log n). (2) If F A k , restricted to length n inputs, can be computed by an unbounded fanin oracle circuit of size s(n) and depth d(n), with k − 1 queries to some set X, then A can be computed with an unbounded fanin (non-oracle) circuit of size n O(k) s(n) and depth d(n) + O(1). (3) Assuming that PH = Σ p 4 ∩ Π p 4 , and < 1, #SAT is not 2 n-enumerable.

Efficient two-dimensional compressed matching

Data Compression Conference, 1992. …, 1992

The problem of two-dimensional run-length compressed matching is defined as follows: Input: Two-d... more

Two-dimensional pattern matching with rotations

Combinatorial Pattern …, 2003

The most efficient currently known algorithms for two dimensional pattern matching with rotations... more The most efficient currently known algorithms for two dimensional pattern matching with rotations have a worst case time complexity of O(n 2 m 3), where the size of the text is n × n and the size of the pattern is m × m. In this paper we present a new algorithm for the problem whose running time is O(n 2 m 2).

Contributions of theoretical computer science

Advanced Query Mechanisms for Biological Databases, Agarwal, Pankaj, Allison, L., Aloy, Patrick,

aaai.org

Advanced Query Mechanisms for Biological Databases,  Agarwal, Pankaj,  Allison, L.,  Aloy, Pa... more Advanced Query Mechanisms for Biological Databases,  Agarwal, Pankaj,  Allison, L.,  Aloy, Patrick,  Altman, Russ B.,  Amir, Amihood,  Ananko, EA,  Atchley, William R.,  Atteson, Kevin,  Automated Clustering and Assembly of Large EST Collections,  Aviles, Francesc X.,  ... Babenko, VN,  Bafna, Vineet,  Baker, Patricia G.,  Baldi, Pierre,  Bayesian Protein Family Classifier,  Bechhofer, Sean,  BioSim—A New Qualitative Simulation Environment for Molecular Biology,  Brass, Andy,  Brunak, ...

Incremental Algorithms for Association Generation

Pacific-Asia Conference on Knowledge Discovery and Data Mining, 1997

Computing Similarity of Run-Length Encoded Strings with Affine Gap Penalty

String Processing and Information Retrieval, 2005

The problem of computing similarity of two run-length encoded strings has been studied for variou... more The problem of computing similarity of two run-length encoded strings has been studied for various scoring metrics. Many algorithms have been developed for the longest common subsequence metric and some algorithms for the Levenshtein distance metric and the weighted edit distance metric. In this paper we consider similarity based on the affine gap penalty metric which is a more general

Fast Parallel and Serial Multidimensional Approximate Array Matching

Sequences, 1990

Improved Deterministic Length Reduction

Computing Research Repository, 2008

This paper presents a new technique for deterministic length reduction. This technique improves t... more This paper presents a new technique for deterministic length reduction. This technique improves the running time of the algorithm presented in (?) for performing fast convolution in sparse data. While the regular fast convolution of vectors V1, V2 whose sizes are N1, N2 respectively, takes O(N1 log N2) using FFT, using the new technique for length reduction, the algorithm proposed

Alphabet independent two dimensional matching

Proceedings of the twenty-fourth annual ACM symposium on Theory of computing - STOC '92, 1992

There are many solutions to the string matching pmZllem whkh are strictly linear in the input siz... more

Optimal two-dimensional compressed matching

Lecture Notes in Computer Science, 1994

Recent proliferation of digitized data and the unprecedented growth in the volume of stored and t... more Recent proliferation of digitized data and the unprecedented growth in the volume of stored and transmitted data motivated the definition of the compressed matching paradigm. This is the problem of efficiently finding a pattern P in a compressed text T without the need to decompress. We present the first optimal two-dimensional compressed matching algorithm. The compression under consideration is the two dimensional run-length compression, used by FAX transmission. We achieve optimal time by proving new properties of two-dimensional periodicity. This enables performing duels in which no witness is required. At the heart of the dueling idea lies the concept that two overlapping occurrences of a pattern in a text can use the content of a predetermined text position or witness in the overlap to eliminate one of them. Finding witnesses is a costly operation in a compressed text, thus the importance of witness-free dueling.

Approximate Matching in Weighted Sequences

Lecture Notes in Computer Science, 2006

Weighted sequences have been recently introduced as a tool to handle a set of sequences that are ... more Weighted sequences have been recently introduced as a tool to handle a set of sequences that are not identical but have many local similarities. The weighted sequence is a “statistical image” of this set, where the probability of every symbol’s occurrence at every text location is given. We address the problem of approximately matching a pattern in such a weighted

Deterministic Length Reduction: Fast Convolution in Sparse Data and Applications

Lecture Notes in Computer Science

In this paper a deterministic algorithm for the length reduction problem is presented. This algor... more In this paper a deterministic algorithm for the length reduction problem is presented. This algorithm enables a new tool for performing fast convolution in sparse data. While the regular fast convolution of vectors V 1 , V 2 whose sizes are N 1 , N 2 respectively, takes O(N 1 log N 2) using FFT, the proposed algorithm performs the convolution in O(n 1 log 3 n 1), where n 1 is the number of non-zero values in V 1. This algorithm assumes that V 1 is given in advance, and the V 2 is given in running time. This running time is achieved using a preprocessing phase on V 1 , which takes O(n 2 1) if N 1 is polynomial in n 1 , and O(n 4 1) if N 1 is exponential in n 1 (which is rarely the case in practical applications). This tool is used to obtain faster results for several well known problems, such as the d-Dimensional Point Set Matching and Searching in Music Archives.

Blocked Pattern Matching Problem and Its Applications in Proteomics

Lecture Notes in Computer Science, 2011

Matching a mass spectrum against a text (a key computational task in proteomics) is slow since th... more Matching a mass spectrum against a text (a key computational task in proteomics) is slow since the existing text indexing algorithms (with search time independent of the text size) are not applicable in the domain of mass spectrometry. As a result, many important applications (e.g., searches for mutated peptides) are prohibitively time-consuming and even the standard search for non-mutated peptides

Efficient randomized dictionary matching algorithms

Combinatorial Pattern Matching, 1992

Abstract The standard string matching problem involves finding all occurrences of a single patter... more Abstract The standard string matching problem involves finding all occurrences of a single pattern in a single text. While this approach works well in many application areas, there are some domains in which it is more appropriate to deal with dictionaries of patterns. A dictionary is a set of patterns; the goal of dictionary matching is to find all dictionary patterns in a given text, simultaneously.