Robert Mateescu

Robert Mateescu
HGST, A Western Digital Company · Research

Ph.D.

About

51
Publications
3,260
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
971
Citations

Publications

Publications (51)
Article
Full-text available
File systems may become corrupted for many reasons despite various protection techniques. Therefore, most file systems come with a checker to recover the file system to a consistent state. However, existing checkers are commonly assumed to be able to complete the repair without interruption, which may not be true in practice. In this work, we demon...
Article
We study the design of storage-efficient algorithms for emulating atomic shared memory over an asynchronous, distributed message-passing network. Our first algorithm is an atomic single-writer multi-reader algorithm based on a novel erasure-coding technique, termed multi-version code. Next, we propose an extension of our single-writer algorithm to...
Article
Rewriting codes can improve the lifetime capacity of NAND flash. All existing q-ary rewriting codes for NAND flash assume that the exact cell levels are known to decoders and thus the number of reads needed is q − 1, while current NAND technology enables low latency page reads by dividing a physical page into log q logical pages so that the average...
Article
Full-text available
Resistive memories, such as phase change memories and resistive random access memories, have attracted significant research interest because of their scalability, non-volatility, fast speed, and rewritability. However, their write endurance needs to be improved substantially for large-scale deployment of resistive memories. In addition, their write...
Conference Paper
Flash memories use the amount of charge (e.g., electrons) trapped in floating gate transistors to represent data. Charge leakage will cause data retention problem by unidirectionally shifting the cell-level distribution. Balanced codes are an effective means to adjust read thresholds adaptively and tolerate the charge leakage under unknown retentio...
Conference Paper
Full-text available
Recently introduced 3D vertical flash memory is expected to be a disruptive technology since it overcomes scaling challenges of conventional 2D planar flash memory by stacking up cells in the vertical direction. However, 3D vertical flash memory suffers from a new problem known as fast detrapping, which is a rapid charge loss problem. In this paper...
Patent
A data storage system includes a memory circuit and a control circuit. The control circuit is operable to receive data bits provided for storage in memory cells of the memory circuit. The control circuit is operable to compare each of the data bits provided for storage in a corresponding one of the memory cells having a stuck-at fault value to the...
Patent
A data storage system includes a memory circuit having memory cells and a control circuit. The control circuit is operable to receive data bits provided for storage in the memory cells. A subset of the memory cells have predetermined stuck-at faults. The control circuit is operable to compute a binomial coefficient for each of the predetermined stu...
Patent
A data storage system includes a memory circuit and a control circuit. The control circuit is operable to receive data bits provided for storage in memory cells of the memory circuit. The control circuit is operable to compare each of the data bits provided for storage in a corresponding one of the memory cells having a stuck-at fault to a value of...
Patent
A data storage system includes a memory circuit comprising memory cells and a control circuit. The control circuit generates a first set of redundant bits indicating bit positions of the memory cells having stuck-at faults in response to a first write operation if a first rate of the stuck-at faults in the memory cells is greater than a first thres...
Patent
Full-text available
A shingled magnetic recording hard disk drive that uses writeable cache tracks in the inter-band gaps between the annular data bands minimizes the effect of far track erasure (FTE) in the boundary regions of annular data bands caused by writing to the cache tracks. Based on the relative FTE effect for all the tracks in a range of tracks of the cach...
Patent
Full-text available
A method and apparatus are provided for implementing enhanced performance for multi-level cell (MLC) memory using threshold-voltage-drift or resistance-drift tolerant moving baseline memory data encoding. A voltage baseline of a prior write is identified, and a data write uses the threshold-voltage-drift or resistance-drift tolerant moving baseline...
Conference Paper
Phase Change Memory (PCM) presents an architectural challenge: writing to it is slow enough to make attaching it to a CPU's main memory controller impractical, yet reading from it is so fast that using it in a peripheral storage device would leave much of its performance potential untapped at low command queue depths, throttled by the high latencie...
Article
Full-text available
Maximum-distance separable (MDS) array codes with high rate and an optimal repair property were introduced recently. These codes could be applied in distributed storage systems, where they minimize the communication and disk access required for the recovery of failed nodes. However, the encoding and decoding algorithms of the proposed codes use ari...
Article
Full-text available
The paper presents an iterative version of join-tree clustering that applies the message passing of join-tree clustering algorithm to join-graphs rather than to join-trees, iteratively. It is inspired by the success of Pearl's belief propagation algorithm as an iterative approximation scheme on one hand, and by a recently introduced mini-clustering...
Article
Full-text available
In Non - ergodic belief networks the posterior belief OF many queries given evidence may become zero.The paper shows that WHEN belief propagation IS applied iteratively OVER arbitrary networks(the so called, iterative OR loopy belief propagation(IBP)) it IS identical TO an arc - consistency algorithm relative TO zero - belief queries(namely assessi...
Article
Full-text available
The paper introduces mixed networks, a new framework for expressing and reasoning with probabilistic and deterministic information. The framework combines belief networks with constraint networks, defining the semantics and graphical representation. We also introduce the AND/OR search space for graphical models, and develop a new linear space searc...
Article
Full-text available
In this paper we compare search and inference in graphical models through the new framework of AND/OR search spaces. Specifically, we com- pare Variable Elimination (VE) and memory- intensive AND/OR Search (AO) and place al- gorithms such as graph-based backjumping and no-good learning, as well as Recursive Condi- tioning (7) and Value Elimination...
Article
Full-text available
Compiling graphical models has recently been under intense investigation, especially for probabilistic modeling and processing. We present here a novel data structure for compiling weighted graphical models (in particular, probabilistic models), called AND/OR Multi-Valued Decision Diagram (AOMDD). This is a generalization of our previous work on co...
Article
Full-text available
In his seminal paper, Pearl [1986] introduced the notion of Bayesian networks and the first processing algorithm, Belief Propagation (BP), that computes posterior marginals, called beliefs, for each variable when the network is singly connected. The paper provided the foundation for the whole area of Bayesian networks. It was the first in a series...
Article
Full-text available
Flash memory is a nonvolatile computer memory comprised of blocks of cells, wherein each cell is implemented as either NAND or NOR floating gate. NAND flash is currently the most widely used type of flash memory. In a NAND flash memory, every block of cells consists of numerous pages; rewriting even a single page requires the whole block to be eras...
Conference Paper
Full-text available
NAND flash memories have become the most widely used type of non-volatile memories. In a NAND flash memory, every block of memory cells consists of numerous pages, and rewriting a single page requires the whole block to be erased. As block erasures significantly reduce the longevity, speed and power efficiency of flash memories, it is critical to m...
Article
Full-text available
The paper investigates parameterized approximate message-passing schemes that are based on bounded inference and are inspired by Pearl's belief propagation algorithm (BP). We start with the bounded inference mini-clustering algorithm and then move to the iterative scheme called Iterative Join-Graph Propagation (IJGP), that combines both iteration a...
Conference Paper
Full-text available
NAND flash memories are the most widely used non-volatile memories, and data movement is common in flash storage systems. We study data movement solutions that minimize the number of block erasures, which are very important for the efficiency and longevity of flash memories. To move data among n blocks with the help of Δ auxiliary blocks, where eve...
Conference Paper
Full-text available
NAND flash memories are currently the most widely used flash memories. In a NAND flash memory, although a cell block consists of many pages, to rewrite one page, the whole block needs to be erased and reprogrammed. Block erasures determine the longevity and efficiency of flash memories. So when data is frequently reorganized, which can be character...
Article
Full-text available
We explore a novel data representation scheme for multi-level flash memory cells, in which a set of n cells stores information in the permutation induced by the different charge levels of the individual cells. The only allowed charge-placement mechanism is a "push-to-the-top" operation which takes a single cell of the set and makes it the top-charg...
Article
Full-text available
Inspired by the recently introduced framework of AND/OR search spaces for graphical models, we propose to augment Multi-Valued Decision Diagrams (MDD) with AND nodes, in order to capture function decomposition structure and to extend these compiled data structures to general weighted graphical models (e.g., probabilistic models). We present the AND...
Article
Full-text available
The paper introduces mixed networks, a new graphical model framework for expressing and reasoning with probabilistic and deterministic information. The motivation to develop mixed networks stems from the desire to fully exploit the deterministic information (constraints) that is often present in graphical models. Several concepts and algorithms spe...
Conference Paper
Full-text available
The paper is an overview of a recently developed compilation data structure for graphical models, with specific application to constraint netw orks. The AND/OR Multi-Valued Decision Diagram (AOMDD) augments well known decision diagrams (OBDDs, MDDs) with AND nodes, in order to capture func- tion decomposition structure. The AOMDD is based on a pseu...
Conference Paper
Full-text available
We propose a new top down search-based algorithm for compiling AND/OR Multi-Valued Decision Diagrams (AOMDDs), as representations ofthe optimal set of solutions for constraint optimization problems. The approach is based on AND/OR search spaces for graphical models, state-of-the-art AND/OR Branch-and-Bound search, and on decision diagrams reduction...
Article
The paper introduces an AND/OR search space perspective for graphical models that include probabilistic networks (directed or undirected) and constraint networks. In contrast to the traditional (OR) search space view, the AND/OR search tree displays some of the independencies present in the graphical model explicitly and may sometimes reduce the se...
Conference Paper
Full-text available
We investigate three parameterized algorithmic schemes for graphical models that can accom- modate trade-offs between time and space: 1) AND/OR Adaptive Caching (AOC(i)); 2) Variable Elimination and Conditioning (VEC(i)); and 3) Tree Decomposition with Conditioning (TDC(i)). We show that AOC(i) is better than the vanilla ver- sions of both VEC(i) a...
Conference Paper
Full-text available
Inspired by AND/OR search spaces for graphical models recently in- troduced, we propose to augment Ordered Decision Diagrams with AND nodes, in order to capture function decomposition structure. This yields AND/OR multi- valued decision diagram(AOMDD) which compiles a constraint network into a canonical form that supports polynomial time queries su...
Conference Paper
Full-text available
In recent work we introduced the perspective of AND/OR search spaces for various types of graphical models (3, 4, 6). In this paper we provide a broader exposition of the subject for the case of constraint networks and we introduce the new concept of semantic width, that captures, more accurately than the well known parameter tree-width, the size o...
Conference Paper
Full-text available
Cutset conditioning is one of the methods of solv- ing reasoning tasks for graphical models, espe- cially when space restrictions make inference (e.g., jointree-clustering) algorithms infeasible. The w- cutset is a natural extension of the method to a hy- brid algorithm that performs search on the condi- tioning variables and inference on the remai...
Conference Paper
Full-text available
In this paper we compare search and inference in graphical models through the new frame- work of AND/OR search. Specifically, we com- pare Variable Elimination (VE) and memory- intensive AND/OR Search (AO) and place al- gorithms such as graph-based backjumping and no-good and good learning, as well as Recur- sive Conditioning (7) and Value Eliminat...
Conference Paper
Full-text available
The contribution of this paper is in vieweing search for constraint processing in the context of AND/OR search spaces and in demonstrating the impact of this view on solutions counting. In a companion paper we introduce the AND/OR search space idea for probabilistic reasoning. In contrast to the traditional (OR) search space view, the AND/OR search...
Article
The paper introduces mixed networks, a new framework for expressing and reasoning with probabilistic and deterministic information. The framework combines belief networks with constraint networks, defining the semantics and graphical representation. We also introduce the AND/OR search space for graphical models, and develop a new linear space searc...
Article
The contribution of this paper is in vieweing search for constraint processing in the context of AND/OR search spaces and in demonstrating the impact of this view on solutions counting. In a companion paper we introduce the AND/OR search space idea for probabilistic reasoning. In contrast to the traditional (OR) search space view, the AND/OR search...
Article
Full-text available
Probabilistic inference in Bayesian networks, and even reasoning within error bounds are known to be NP-hard problems. Our research focuses on investigating approximate message-passing algorithms inspired by Pearl's belief propagation algorithm and by variable elimination. We study the advantages of bounded inference provided by anytime schemes suc...
Article
Full-text available
The paper presents an iterative version of join-tree clustering that applies the message passing of join-tree clustering algorithm to join-graphs rather than to join-trees, iteratively.
Article
Full-text available
The paper presents a parameterized approximation scheme for probabilistic inference. The scheme, called Mini-Clustering (MC), extends the partition-based approximation offered by mini-bucket elimination, to tree decompositions. The benefit of this extension is that all single-variable beliefs are computed (approximately) at once, using a two-phase...
Article
Full-text available
The paper presents a parameterized approximation scheme for probabilistic inference. The scheme, called Mini-Clustering (MC) extends the partition-based approximation offered by mini-bucket elimination, to tree decompositions. The benefit of this extension is that all single variable beliefs are computed (approximately) at once, using a twophase me...
Article
Full-text available
The Boolean Satisfiability (SAT) solvers have been in recent years very successful at solving very large practical prob-lems, with hundreds of thousands of variables and millions of clauses. Although SAT can be viewed as a graphical model, it is intriguing that SAT solvers do not seem to perform any-where close to the worst case upper bound that is...
Article
Full-text available
We investigate two parameterized algorithmic schemes for graphical models that can accommodate trade-offs between time and space: 1) AND/OR Cutset Conditioning (AOC(i)) and 2) Variable Elimination with Conditioning (VEC(i)). We show that AOC(i) is better than the vanilla versions of VEC(i), and use the guiding principles of AOC(i) to improve VEC(i)...