forked from a3d3-institute/a3d3-institute.github.io
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathhac.html
133 lines (108 loc) · 5.18 KB
/
hac.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
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
<!DOCTYPE HTML>
<!--
Editorial by HTML5 UP
html5up.net | @ajlkn
Free for personal and commercial use under the CCA 3.0 license (html5up.net/license)
-->
<html>
<head>
<title>A3D3</title>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1, user-scalable=no" />
<link rel="stylesheet" href="assets/css/main.css" />
</head>
<body class="is-preload">
<!-- Wrapper -->
<div id="wrapper">
<!-- Main -->
<div id="main">
<div class="inner">
<!-- Header -->
<header id="header">
<a href="index.html" class="logo"><strong>A3D3</strong></a>
<ul class="icons">
<li><a href="https://twitter.com/a3d3institute" class="icon brands fa-twitter"><span class="label">Twitter</span></a></li>
</ul>
</header>
<!-- Content -->
<section>
<header class="main">
<h1>Hardware and Algorithm Codevelopment</h1>
</header>
<!--
<span class="image main"><img src="images/pic11.jpg" alt="" /></span>
-->
<p>
Developing AI methods to encode non-lattice-structured
data is one main challenge in current AI systems. Such data structures are common in the scientific applications to be explored in A3D3. AI researchers have to carefully consider the specifics of the data irregularity, the chosen model architecture, and the system constraints. A3D3 proposes to investigate optimal encoders for these non-lattice-structured scientific data.
Within the A3D3 science drivers, fast identification of sporadic events from large-scale data streams provides a unique opportunity for discovery.Traditional selection strategies use either subjective heuristics (e.g., a high energy collision or a bright transient) or model-based classifiers that are trained with specific labels. Such strategies exclude unconventional signatures with unexpected properties where manually defined heuristics and labeled examples are insufficient to explain them. Decoders based on encoded data can be designed to overcome these challenges. A3D3 will explore semi-supervised/unsupervised learning methods for e↵ective anomaly, or out-of-class, detection. Recent work on variational autoencoders (VAEs) and generative models learn the distribution of the input data and detect anomalies based on their likelihood scores.
Hardware-AI co-design is the tight coupling of the design of AI algorithms with the domain science, hardware, and system constraints.
A hallmark of co-design is either an encoding of the design constraints directly in the cost function of the training algorithm or a fast-to-evaluate approximation of the acceptability of an on-device solution to allow for fast exploration and evaluation.
Our primary focus is on achieving low latency, real-time processing of scientific data. A3D3 aims to design end-to-end workflows integrated with user-friendly tools for hardware-aware co-design of AI algorithms.
</section>
</div>
</div>
<!-- Sidebar -->
<div id="sidebar">
<div class="inner">
<!-- Search -->
<section id="search" class="alt">
<form method="post" action="#">
<input type="text" name="query" id="query" placeholder="Search" />
</form>
</section>
<!-- Menu -->
<nav id="menu">
<header class="major">
<h2>Menu</h2>
</header>
<ul>
<li><a href="index.html">Home</a></li>
<li><a href="about.html">About</a></li>
<li>
<span class="opener">Team</span>
<ul>
<li><a href="researchers.html">Researchers</a></li>
<li><a href="trainees.html">Trainees</a></li>
<li><a href="institutions.html">Institutions</a></li>
<li><a href="ExecBoard.html">Executive Board</a></li>
<li><a href="advisory.html">Advisory board</a></li>
<li><a href="collaborators.html">Collaborators</a></li>
</ul>
</li>
<li>
<span class="opener">Activities</span>
<ul>
<li><a href="hac.html">HAC</a></li>
<li><a href="hep.html">HEP</a></li>
<li><a href="mma.html">MMA</a></li>
<li><a href="neuros.html">Neuroscience</a></li>
<li><a href="publications.html">Publications</a></li>
</ul>
</li>
<li>
<span class="opener">Connect</span>
<ul>
<li><a href="events.html">Events</a>
<li><a href="news.html">News</a>
<li><a href="contact.html">Contact</a></li>
</ul>
</li>
<li><a href="jobs.html">Jobs</a></li>
</ul>
</nav>
<!-- Footer -->
<footer id="footer">
<p class="copyright">© A3D3. All rights reserved. Design: <a href="https://html5up.net">HTML5 UP</a>.</p>
</footer>
</div>
</div>
</div>
<!-- Scripts -->
<script src="assets/js/jquery.min.js"></script>
<script src="assets/js/browser.min.js"></script>
<script src="assets/js/breakpoints.min.js"></script>
<script src="assets/js/util.js"></script>
<script src="assets/js/main.js"></script>
</body>
</html>