-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
executable file
·111 lines (91 loc) · 5.65 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="description" content="">
<meta name="author" content="">
<title>Ravishankar Krishnaswamy</title>
<!-- Bootstrap Core CSS -->
<link href="bootstrap.min.css" rel="stylesheet">
<!-- Custom CSS -->
<style>
body {
padding-top: 70px;
/* Required padding for .navbar-fixed-top. Remove if using .navbar-static-top. Change if height of navigation changes. */
}
</style>
<!-- HTML5 Shim and Respond.js IE8 support of HTML5 elements and media queries -->
<!-- WARNING: Respond.js doesn't work if you view the page via file:// -->
<!--[if lt IE 9]>
<script src="https://oss.maxcdn.com/libs/html5shiv/3.7.0/html5shiv.js"></script>
<script src="https://oss.maxcdn.com/libs/respond.js/1.4.2/respond.min.js"></script>
<![endif]-->
</head>
<body>
<!-- Navigation -->
<nav class="navbar navbar-inverse navbar-fixed-top" role="navigation">
<div class="container">
<!-- Brand and toggle get grouped for better mobile display -->
<div class="navbar-header">
<button type="button" class="navbar-toggle" data-toggle="collapse" data-target="#menu">
<span class="sr-only">Toggle navigation</span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
</button>
<a class="navbar-brand" href="#">Ravishankar Krishnaswamy</a>
</div>
<!-- Collect the nav links, forms, and other content for toggling -->
<div class="collapse navbar-collapse" id="menu">
<ul class="nav navbar-nav navbar-right">
<li style="background-color:rgba(65,65,65,1);">
<a href="index.html">Home</a>
</li>
<li>
<a href="pubs.html">Publications</a>
</li>
<li>
<a href="courses.html">Teaching</a>
</li>
<li>
<a href="misc.html">Miscellaneous</a>
</li>
</ul>
</div>
<!-- /.navbar-collapse -->
</div>
<!-- /.container -->
</nav>
<!-- Page Content -->
<div class="container">
<div class="row">
<br>
<p>I am a principal researcher at <a href="https://www.microsoft.com/en-us/research/lab/microsoft-research-india/">Microsoft Research India</a>. I completed my PhD at <a href="http://www.cs.cmu.edu">Carnegie Mellon University</a> in 2012 where I was fortunate to have <a href="http://www.cs.cmu.edu/~anupamg">Anupam Gupta</a> as my advisor. From 2012-2014, I was a <a href="https://www.simonsfoundation.org/">Simons Postdoctoral Fellow</a> at the CS Department in <a href="http://www.cs.princeton.edu">Princeton University</a>, and I also spent Fall 2014 visiting the <a href="http://ieor.columbia.edu/">IEOR Department </a> in Columbia University. Long ago, I was an undergrad at <a href="http://www.cse.iitm.ac.in">IIT Madras</a>.
</p>
<p>My recent research has been on designing new algorithms and data structures for very large-scale vector search a.k.a approximate nearest neighbor search. With the emergence of deep learning, almost all information is represented as vectors in high-dimensional space, and almost most search problems crucially depend on being able to identify <i>nearby vectors</i> (corresponding to the base corpos of information) to the query's vector embedding.
To ensure scalability -- which was the limiting factor prior to our work -- the system we have designed, called DiskANN, is a hybrid data-structure which stores only a small fraction of the data structure in main memory, and the bulk of the index resides on cheaper but slower SSDs. The crucial technical challenge is in ensuring that our search methods don't access the SSD-portion of the index very often due to the long latencies. DiskANN has been used extensively within Microsoft, both for first-party use-cases as well as third-party offerings through Azure.
<br><br>
I am also interested in approximation algorithms, especially for graph-(connectivity/flow) problems and clustering problems, and in studying models which incorporate <i>uncertainty in the input</i> such as online algorithms and stochastic optimization. <a href="http://repository.cmu.edu/dissertations/157/">Here is a link</a> to my PhD thesis on <b>Approximation Techniques for Stochastic Combinatorial Optimization</b>. </p>
</div>
<hr>
<div class="row">
<b>Contact</b>
</div>
<div class="row">
<div class="col-md-6" style="padding:0; margin:0; border:0;">Vigyan Building, 9, Lavelle Road, Bangalore 560018</div>
<div class="col-md-6" style="text-align: right;padding:0; margin:0; border:0;">ravishan (at) cs (dot) cmu (dot) edu</div>
</div>
<!-- /.row -->
</div>
<!-- /.container -->
<!-- jQuery Version 1.11.1 -->
<script
src="https://code.jquery.com/jquery-3.1.1.min.js"
integrity="sha256-hVVnYaiADRTO2PzUGmuLJr8BLUSjGIZsDYGmIJLv2b8="
crossorigin="anonymous"></script>
<!-- Bootstrap Core JavaScript -->
<script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/js/bootstrap.min.js"></script>
</body>
</html>