CSE 254 Handout
Nearest Neighbor Preserving Embeddings
(paper by Piotr Indyk and Assaf Naor) Konstantin Pervyshev [email protected]
1
Nearest Neighbor Search
Problem 1 (The nearest neighbor problem). Given a set X Rn , build a data structure which giv
Dept_name
Location
Department
Headed by
Offers
HAS
course_name
Is taught by
duration
Course
Instructor
Course_id
byby
Enrolled by
First Name
Instructor_id
b
phone
Last name
Student
Last_name
phone
Student_id
First_name
#include<iostream>
#include<exception>
using namespace std;
int. main()
cfw_
double a,b;
cout<"enter any two numbers "<endl;
cin>a>b;
try
cfw_
if(b=0)
throw "Division by zero";
cout<"devision is : "<a/b<endl;
catch(char*c)
cfw_
cout<"exception due to "<c<
`
What is journalist ?
Ans. In professional news media, all reporters are journalists but not all journalists are
reporters. "Journalist" includes anyone who is part of the editorial process of gathering and
disseminating news: Reporters, editors, produce
Distance
Distance Scales,
Embeddings, and Metrics of
Negative Type
By James R. Lee
Presented by Andy Drucker
Mar. 8, 2007
CSE 254: Metric Embeddings
Negative Type Metrics
Negative Type Metrics
Also, L_1 is NEG
Finally, NEG metrics arise in SDP instances f
Distortion Bounds from Coloring Finding the Coloring
Bounded Geometries, Fractals, and Low-Distortion Embeddings[1]
Anupam Gupta, Robert Krauthgamer and James R. Lee
February 27, 2007
presented by Brian McFee for CSE254
Distortion Bounds from Coloring Fin
Measured Descent: A new embedding method for nite metrics
R. Krauthgamer, J. Lee, M. Mendel, A. Naor (2004)
Metric Embeddings Seminar
March 6, 2007
Presentation by Evan Ettinger
KLMN04 (Metric Embeddings Seminar)
Measured Descent
March 6, 2007
1 / 22
The
CSE 254: Seminar on Learning Algorithms: Embeddings
March 13th, 2007
Dimension reduction in L1 : a negative result
Professor: Sanjoy Dasgupta Lecturer: Claire Monteleoni
The Johnson-Lindenstrauss lemma shows that only d = O( 1 log n) dimensions are needed
Small distortion and volume preserving embedding for Planar and Euclidian metrics
Satish Rao
presented by Fjola Run Bjornsdottir CSE 254 - Metric Embeddings Winter 2007
CSE 254 - Metric Embeddings - Winter 2007 p. 1/2
Overview
Denitions Main Results Proo
Nearest Neighbor Searching and Metric Space Dimensions
Presented by: Nakul Verma Feb 8, 2007.
Outline
Motivation Introduction to various concepts of dimensions Applications to Nearest Neighbor Queries
2
Outline
Motivation Introduction to various concepts
Approximating finite metrics by distributions over tree metrics
Daniel Hsu, CSE 254
1 Introduction
We'd like to approximate arbitrary finite metric spaces (X, ) with simpler ones. So far, the approximating (host) metric spaces we've considered have been (
Multicommodity Flows Handout
Roy Liu Winter 2007
Handout
Multicommodity Flows
February 6, 2007
1
Background
In the traditional (s, t) flow setting, we want to push as many units of a single resource, or commodity, from the source s to the sink t. Upon gen
N OTES
ON LOCALITY SENSITIVE HASHING WITH
p- STABLE DISTRIBUTIONS
L. Cayton
We look at a hash scheme used to compute approximate nearest neighbors quickly in large, high-dimensional databases.
1
p-Stable distributions
Before turning to the nearest neighbo