The Smart Access Memory: An Intelligent RAM for Nearest Neighbor Database Searching
نویسنده
چکیده
An Intelligent RAM for Nearest Neighbor Database Searching Aaron Lipman and Woodward Yang Abstract The nearest neighbor algorithm is well suited to an IRAM implementation, given its high bandwidth requirement between memory and processing. We have designed and implemented the Smart Access Memory (SAM) chip to retrieve the k nearest neighbors to a query point in a database of example data vectors by integrating 64 bitserial processors, 4096 sorting units, and 16Mb of memory using a 0.35 m commercial DRAM process (the 1.6mm 1.6mm chip is currently in fabrication.) The 800K transistor processor array can be clocked at speeds between 100200MHz providing an internal memory bandwidth of 6.412.8Gb/s. This approach also allows for multiple chips to be used in parallel, and a system of 16 SAM chips can accelerate a brute-force nearest neighbor search by over three orders of magnitude compared to a highly optimized software implementation on a 100MHz PA-RISC workstation.
منابع مشابه
Non-zero probability of nearest neighbor searching
Nearest Neighbor (NN) searching is a challenging problem in data management and has been widely studied in data mining, pattern recognition and computational geometry. The goal of NN searching is efficiently reporting the nearest data to a given object as a query. In most of the studies both the data and query are assumed to be precise, however, due to the real applications of NN searching, suc...
متن کاملAn Improved K-Nearest Neighbor with Crow Search Algorithm for Feature Selection in Text Documents Classification
The Internet provides easy access to a kind of library resources. However, classification of documents from a large amount of data is still an issue and demands time and energy to find certain documents. Classification of similar documents in specific classes of data can reduce the time for searching the required data, particularly text documents. This is further facilitated by using Artificial...
متن کاملAn Improved K-Nearest Neighbor with Crow Search Algorithm for Feature Selection in Text Documents Classification
The Internet provides easy access to a kind of library resources. However, classification of documents from a large amount of data is still an issue and demands time and energy to find certain documents. Classification of similar documents in specific classes of data can reduce the time for searching the required data, particularly text documents. This is further facilitated by using Artificial...
متن کاملElderly Daily Activity-Based Mood Quality Estimation Using Decision-Making Methods and Smart Facilities (Smart Home, Smart Wristband, and Smartphone)
Due to the growth of the aging phenomenon, the use of intelligent systems technology to monitor daily activities, which leads to a reduction in the costs for health care of the elderly, has received much attention. Considering that each person's daily activities are related to his/her moods, thus, the relationship can be modeled using intelligent decision-making algorithms such as machine learn...
متن کاملImproving Bilayer Product Quantization for Billion-Scale Approximate Nearest Neighbors in High Dimensions
The top-performing systems for billion-scale high-dimensional approximate nearest neighbor (ANN) search are all based on two-layer architectures that include an indexing structure and a compressed datapoints layer. An indexing structure is crucial as it allows to avoid exhaustive search, while the lossy data compression is needed to fit the dataset into RAM. Several of the most successful syste...
متن کامل