aboutsummaryrefslogtreecommitdiff
path: root/20220816research.txt
blob: 65208f2cd7084ad301f2bcce4c516a6d0d91904e (plain)
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
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
Niti's Research Prof: frank.dellaert@cc.gatech.edu

Do Undergrad Research if:
- Want more depth in a specific area not covered in class
- Want to look at research for PhD

Don't Undergrad Research if:
- You want certain results
- Want individual attention from faculty
- Want recommendations for jobs (faculty recs don't go very far in
  industry anyway)

That fits me!

Course Reg Numbers: [2/4]69[8/9]
2: fresh/soph
4: jun/sen
8: paid
9: unpaid
|Sometimes not required but highly recommended

Meets the junior design requirement
- LMC 4701/2
- 9 credits of undergrad research (2 sem unpaid iirc)

BS/MS apply after 60 credits; can do research as project/course option

Undergrad research can be "academic" or "implementation-oriented."
- Much less grant- and paper-oriented than Masters/grad research.

# Hyesoon Kim
## Computer Architecture
Working on GPU processor design
- CUDA Everywhere
- Open Source Design
- CuPBoP??

Looking for C/C++ skill set
GPU (CUDA/OpenCL) programmer
LLVM experience hopefully
Interested in compiler, runtime system

# Frank Li
## Data-Driven Methods for Improving Internet Security

BEES Lab.
Applies "empirical approaches" to Internet security concerns.
Probably a solutions-y thinktank?
- Large-scale Internet/Web Measurements
    - Ex: massscan IPv4 range (monitor server mgmt behaviors)
    - Develop new techniques for IPv6 internet
- Also data mines auth mechanisms & password leaks
- Analyze open-source Git repos to understand secure software
  development
- Characterize passive DNS data to understand user network traffic
  patterns

Doesn't feel like it's for me. Too much of a social-science study

- Qualitative User Studies
    - Understand how sysadmins manage software updates
    - Understand webadmins password authentication policies
    - How users react to security notices/warnings

Expects background in web security I think?

# Rogues Gallery VIP Team
## What?
Transistor scaling/hitting the end of Moore's law.
Investigates new tech for continuing to improve performance

"Rogues Gallery Testbed" (FPGA & FPAA)
- Unique, "rogue" hardware that needs to be programmed into useful
  applications---unlike commodity memory/processors

Worked on:
neuromorphic something
emu chick neo-memory
quantum computing
maybe reconfigurable computing

Needs skills:
- Min: knowledge of linux, ssh, programming (C/C++/Pyth), lin algebra
- Preferred: ECE 3400 (FPAA), CX 4220, and another class I haven't taken

# Karen Feigh
## UROC Opportunities with Helicopters

- Cognitive Engineering
    - What if we designed training and technology with how people
      actually do their work?
- procedures and displays for air and space
- basic research for mil command & control decisionmaking
- adaptive intelligence for industrial robots & autonomous vehicles
- Examine human-machine interaction using field work, math, etc.

Hiring into projects (multiple year projects getting started):
- Humans & AI Piloted Helicopter (special forces/medevac)
- Helicopter Multi-modal Symbology
Working on better AI/better human interfacing

Requesting:
- Experience w/ unreal engine

# Wei Xu
## NLP Research Lab

NLP side
- Text generation
- Human-centered eval
- Multilingual NLP
- Stylistics

Machine Learning
- Sequence-to-sequence models
- Controllability of neural network
- Cross-lingual transfer learning
- Zero-shot/few-shot learning
- Robust domain adaptation

Want possible multilingual, interested in linguistics.
Didn't mention ML proficiency already.

seq2seq: simplify english
Social Media Data: like hashtag word analysis

# Anirudh Rama (or the actual prof?)
## Community wireless networks/organic mesh networks

- Affordable internet connection
- Networks hosting local services
- Disaster scenarios

Studying:
- Service performance/putting servers on mesh
- Handling unpredictable network topology like wireless links,
  obstructions, unreliable nodes, difficult issue diagnosis

Setting up a mesh testbed on GT campus (wireless router/antenna on
buildings; monitor perf; video conference server)

Preferred: CS 2200, Linux SSH, Basic Networking

# Vivek Sarkar
## Habanero Extreme Scale Software Research Lab

- Extreme parallelism/heterogeneity at the end of Moore's Law
- Distributed parallelism, new software foundation of compilers,
  runtimes, debuggers, programming languages, etc.
- Parallelism cliffs are very hard for something like Python code.

!! See orthogonal persistence techniques & ideas in Houyhhnum Computing

- Programming models (like better/safer primitives in C++, Pyth, Rust)
- Compilers
- Runtime Systems
- Handling Debugging Better

Working on async bulk-synchronous actor-model parallelism

1 Interested in graph algo, distributed machine learning
    - Rust/C++ actor-based programming model
2 Pyth kernels using Intrepydd optimizing compiler
    - Sparse matrix/tensor computation, pandas
3 Statistic & Dynamic Program Analysis Techniques

# Florian Schaefer
## Inference and Computation|Competitive Optimization
- Inf & Comp
Conditional (normal?) distributions have a very deep connection to
factorizing matrices with Cholesky.
Creates fast algorithms for stats & physical simulations

- Competitive Optimization
Sort of game theory, adversarial learning, and other.
Gradient Descent? Competitive Gradient Descent
    - Machine Learning for PDEs

Very cool! Very mathy!!!

Interested in:
- Strong math affinity
- Programming experience with lots of new frameworks
- Likely 1-on-1 meetings
- Expects substantial effort on weekly basis
- Wants you to become published in a good venue (~1-1.5yrs)

# Spencer Bryngelson
## Fluid dynamics for humans

Algorithms for flows (mechanics/fluid dynamics)
- Computational Methods

He works in the computational physics/supercomputing research group.
Industrial-scale simulations!!! (Like exascale/perfect scaling on a big
computer)

Wants:
- Intermediate expertise in C/C++ or Fortran
- Strong handle on numerical linear algebra
- Knowledge of numerics for PDEs
Interested in:
- Computing at huuuge scale
- Using latest HPC hardware
- Solving pressing problems in health/defense

# Thad Starner
## Contextual Computing Group

Works on Google Glass woah.
Product originator, still works on new stuff.
Head Worn Display.

- One research in sign language recognition stuff
    - Significant problem for language abilities of deaf children
    - Using MediaPipe for widescale sign-language recognition on a
      smartphone
- Cetacean Hearing & Telemetry
- Research Group also studies: brain-neural interfacing

Wants:
- Sophomores & Freshmen for long-term study
- Send email with UROC + Name + Project Interest to thad@gatech

Not really my thing.

Running List
------------
- Hyesoon Kim (GPU/EE stuff)
- Rogues Gallery (but only the research side pls)
- Wei Xu (NLP)
- Anirudh Rama (community mesh network topology)
- Vivek Sarkar (Habanero) - somewhat cool but outside my wheelhouse
- Florian Schaefer (mathy guy) - yes yes yes
- Spencer Bryngelson (fluid dynamics & hpc)

Possibly
- Niti's professor
- any other HPC researchers?