-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
356 lines (335 loc) · 13.9 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
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
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
---
layout: default
title: Home
menu_title: Home
menu_weight: 1
---
<br>
<div class='d-md-flex justify-content-center'>
<div class='p-2 flex'>
<img src='assets/profile.jpg' class="profile"/>
</div>
<div class='p-2 flex'>
<br>
<font size="+1"><strong>Aaditya Naik</strong></font><br>
PhD Student <br>
<a href="http://www.cis.upenn.edu">Computer and Information Science</a><br>
<a href="http://www.upenn.edu">University of Pennsylvania</a><br>
Email: <a href="mailto:asnaik@seas.upenn.edu" target="_blank">asnaik@seas.upenn.edu</a><br>
<span class="center">
<a href="assets/CV.pdf" class="icon-link" style="padding-left: 0;">C.V.</a>
<!-- | -->
<a href="https://www.github.com/aadityanaik" class="icon-link" target="_blank"><i class="fab fa-github icon"></i></a>
<!-- | -->
<a href="https://www.linkedin.com/in/aaditya-naik" class="icon-link" target="_blank"><i class="fab fa-linkedin icon icon-li"></i></a>
<!-- | -->
<a href="https://twitter.com/aaditya_naik" class="icon-link" target="_blank"><i class="fab fa-twitter icon icon-li"></i></a>
<!-- | -->
<!-- <a href="https://scholar.google.com/citations?user=EfE0jh4AAAAJ&hl=en" class="icon-link" target="_blank"><i class="fab fa-google-scholar icon icon-li"></i></a> -->
</span>
</div>
</div>
<!--
<h3><center><b><a href="faq.html">Information about CIS 450/550 in Spring 2023</a></b></center></h3>
-->
<hr style="color: black; height: 1px; background-color:black;" />
<div class="row">
<div class="col-md-12" id="about">
<h3>About Me</h3>
</div>
</div>
<p> Welcome to my corner of the internet! I am a 5th year PhD student at the <a href="https://www.upenn.edu">University of Pennsylvania</a>
advised by Prof. <a href="https://www.cis.upenn.edu/~mhnaik/">Mayur Naik</a>.</p>
<p>
My research interests span Programming Languages and Machine Learning.
The overarching goal of my research is to enable machine learning practitioners enable machine learning practitioners to write
neurosymbolic programs that can scale to highly complex tasks and datasets using intuitive and expressive interfaces.
My work focuses on combining symbolic reasoning with deep learning to address tasks requiring both perception and
logic-based reasoning, while ensuring scalability through techniques like vectorized computations and GPU-accelerated differentiable reasoning.</p>
<p>
I am also interested in developing techniques to help machine learning practitioners effectively
understand where their models fail and identify ways to fix them. My other research interests include designing program
synthesis techniques to streamline software analysis, bug finding, and code generation, as well as exploring compiler-based optimization
strategies for large-scale machine learning frameworks in challenging domains such as image processing, text understanding,
and multimodal reasoning.</p>
<hr style="color: black; height: 1px; background-color:black;" />
<div class="row">
<div class="col-md-12" id="news">
<h3>News</h3>
<ul>
<li>
I presented our paper "TorchQL: A Programming Framework for Integrity Constraints in Machine Learning" in
the proceedings of <span style="font-weight: bold;">OOPSLA 2024</span>.
Try out TorchQL <a href="https://github.com/TorchQL/torchql" target="_blank">here</a>, and read our
paper <a href="https://arxiv.org/abs/2308.06686" target="_blank">here</a>.
</li>
<li>
Our work "LLM Based Test-Driven Interactive Code Generation: User Study and Empirical Evaluation" is published
in the <span style="font-weight: bold;">IEEE Transactions on Software Engineering</span>. Read the article
<a href="https://ieeexplore.ieee.org/document/10606356" target="_blank">here</a>.
</li>
<li>
I am very grateful to be awarded the 2023 <span style="font-weight: bold;">Google PhD Fellowship</span>
in Programming Technology and Software Engineering.
</li>
</ul>
</div>
</div>
<hr style="color: black; height: 1px; background-color:black;" />
<div class="row">
<div class="col-md-12" id="research">
<h3>Research</h3>
<p>
While machine learning has seen several advances in recent years, with models achieving state-of-the-art performance
on a variety of tasks, analyzing and understanding these models and their failures is an ad-hoc and often chaotic process.
This is exacerbated by the lack of tools and frameworks that allow practitioners to interactively explore their models
in a manner that is intuitive and easily accessible.
My research aims to bridge this gap by developing novel techniques and tools to allow the systemic analysis
and debugging of machine learning models.
</p>
<p>
<h4>TorchQL</h4>
One of the key challenges in analyzing machine learning models is the lack of a uniform interface to interact with them.
While there are several tools that allow practitioners to analyze their models, these tools are often limited in scope
and are not easily extensible to new models or tasks.
To address this, I am developing a novel querying language called <a href="https://github.com/TorchQL/torchql">TorchQL</a>
that allows practitioners to directly query their models and datasets in a style akin to querying frameworks
like MongoDB and SQL. This allows practitioners to craft intricate and complex queries that characterize the
errors in their models and can generalize to identify similar errors in unseen data as well as in other models.
These queries can identify a range of issues, from simple classification errors to violations of domain
knowledge, biases and labeling errors in the training data, distribution shift, and more.
Read more about TorchQL in our paper <a href="https://arxiv.org/abs/2308.06686">here</a>.
</p>
<p>
<h4>SQRL</h4>
In addition to TorchQL, my framework SQRL (pronounced <span style="font-style: italic;">squirrel</span>) uses
data-driven program synthesis techniques to characterize the errors in machine learning models in terms of
grounded concepts and relations intuitive to practitioners.
You can read more about SQRL in our blog post <a href="https://debugml.github.io/SQRL/" target="_blank">here</a>.
</p>
<p>
<h4>Program Synthesis</h4>
I am also interested in using machine learning to develop program synthesis techniques, specifically
Inductive Logic Programming (ILP) and applying them to a variety of tasks, including software analysis,
bug finding, and code generation.
</p>
<p>
If any of this interests you, I am actively looking for collaborators and would love to chat!
Feel free to reach out to me by email <a href="mailto:asnaik@seas.upenn.edu" target="_blank">here</a>.
</p>
</div>
</div>
<hr style="color: black; height: 1px; background-color:black;" />
<div class="row">
<div class="col-md-12" id="publications">
<h3>Publications</h3>
</div>
</div>
<div class="row">
<div class="col-md-12 pubdiv" >
<h4>Recent Manuscripts</h4>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">Dolphin: A Programmable Framework for Scalable Neurosymbolic Learning</h5>
<div class="pubauths">
<span style="font-weight: bold">Aaditya Naik*</span>, Jason Liu, Claire Wang, Saikat Dutta, Mayur Naik, Eric Wong
</div>
<div class="publinks">
Arxiv Preprint.
[
<a href="https://arxiv.org/pdf/2410.03348" target="_blank">Paper</a>
]
</div>
</div>
</div>
<br>
<div class="row">
<div class="col-md-12 pubdiv" >
<h4>Conference Papers</h4>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">TorchQL: A Programming Framework for Integrity Constraints in Machine Learning</h5>
<div class="pubauths">
<span style="font-weight: bold">Aaditya Naik</span>, Adam Stein, Yinjun Wu, Mayur Naik, Eric Wong
</div>
<div class="publinks">
Conditionally accepted to <span style="font-weight: bold;">OOPSLA 2024</span>.
<!-- <br> -->
[
<a href="https://arxiv.org/abs/2308.06686" target="_blank">Paper</a>
],
[
<a href="https://github.com/TorchQL/torchql" target="_blank">Code</a>
],
[
<a href="https://colab.research.google.com/drive/1dXsyx20GK6OXuRsQzwANlZzu_0mFqtrZ" target="_blank">Demo</a>
]
</div>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">Relational Query Synthesis ⨝ Decision Tree Learning</h5>
<div class="pubauths">
<span style="font-weight: bold">Aaditya Naik</span>, Aalok Thakkar, Adam Stein, Mayur Naik, Rajeev Alur
</div>
<div class="publinks">
Proceedings of <span style="font-weight: bold;">VLDB 2024</span>.
<!-- <br> -->
[
<a href="assets/papers/vldb24_libra.pdf" target="_blank">Paper</a>
]
</div>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">Do Machine Learning Models Learn Statistical Rules Inferred from Data?</h5>
<div class="pubauths">
<span style="font-weight: bold">Aaditya Naik</span>, Yinjun Wu, Mayur Naik, Eric Wong
</div>
<div class="publinks">
Proceedings of <span style="font-weight: bold;">ICML 2023</span>.
<!-- <br> -->
[
<a href="https://arxiv.org/pdf/2303.01433.pdf" target="_blank">Paper</a>
]
</div>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">CodeTrek: Flexible Modeling of Code using an Extensible Relational Representation</h5>
<div class="pubauths">
Pardis Pashakhanloo, <span style="font-weight: bold">Aaditya Naik</span>, Yuepeng Wang, Hanjun Dai, Petros Maniatis, Mayur Naik
</div>
<div class="publinks">
Proceedings of <span style="font-weight: bold;">ICLR 2022</span>.
<!-- <br> -->
[
<a href="https://openreview.net/pdf?id=WQc075jmBmf" target="_blank">Paper</a>
],
[
<a href="https://openreview.net/forum?id=WQc075jmBmf" target="_blank">Open Review Forum</a>
],
[
<a href="https://github.com/ppashakhanloo/CodeTrek" target="_blank">Code</a>
]
</div>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">Sporq: An Interactive Environment for Exploring Code Using Query-by-Example</h5>
<div class="pubauths">
<span style="font-weight: bold">Aaditya Naik</span>, Jonathan Mendelson, Nathaniel Sands, Yuepeng Wang, Mayur Naik, Mukund Raghothaman
</div>
<div class="publinks">
Proceedings of <span style="font-weight: bold;">UIST 2021</span>.
<!-- <br> -->
[
<a href="assets/papers/uist21_sporq.pdf" target="_blank">Paper</a>
],
[
<a href="assets/videos/demo_video.mp4" target="_blank">Demo</a>
]
</div>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">Example-Guided Synthesis of Relational Queries</h5>
<div class="pubauths">
Aalok Thakkar, <span style="font-weight: bold">Aaditya Naik</span>, Nate Sands, Mukund Raghothaman, Mayur Naik, Rajeev Alur
</div>
<div class="publinks">
Proceedings of <span style="font-weight: bold;">PLDI 2021</span>.
<!-- <br> -->
[
<a href="assets/papers/pldi21_egs.pdf" target="_blank">Paper</a>
],
[
<a href="https://github.com/aalok-thakkar/egs-artifact" target="_blank">Code</a>
]
</div>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">GenSynth: Synthesizing Datalog Programs without Language Bias</h5>
<div class="pubauths">
Jonathan Mendelson*, <span style="font-weight: bold">Aaditya Naik*</span>, Mukund Ragothaman, Mayur Naik
</div>
<div class="publinks">
Proceedings of <span style="font-weight: bold;">AAAI 2021</span>.
<!-- <br> -->
[
<a href="assets/papers/aaai21_gensynth.pdf" target="_blank">Paper</a>
],
[
<a href="assets/posters/aaai21_gensynth_poster.pdf" target="_blank">Poster</a>
],
[
<a href="https://jonomendelson.github.io/gensynth/" target="_blank">Website</a>
],
[
<a href="git@github.com:jonomendelson/gensynth.git" target="_blank">Code</a>
]
</div>
</div>
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">Code2Inv: A Deep Learning Framework for Program Verification</h5>
<div class="pubauths">
Xujie Si*, <span style="font-weight: bold">Aaditya Naik*</span>, Hanjun Dai, Mayur Naik, Le Song
</div>
<div class="publinks">
Proceedings of <span style="font-weight: bold;">CAV 2020</span>.
<!-- <br> -->
[
<a href="assets/papers/cav20_code2inv.pdf" target="_blank">Paper</a>
],
[
<a href="assets/Slides/CAV20_122_slides.pptx" target="_blank">Slides</a>
],
[
<a href="code2inv.cis.upenn.edu/" target="_blank">Website</a>
],
[
<a href="https://github.com/PL-ML/code2inv" target="_blank">Code</a>
]
</div>
</div>
</div>
<br>
<div class="row">
<div class="col-md-12 pubdiv" >
<h4>Journal Publications</h4>
</div>
<div class="row">
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">LLM-Based Test-Driven Interactive Code Generation: User Study and Empirical Evaluation</h5>
<div class="pubauths">
Sarah Fakhoury, <span style="font-weight: bold">Aaditya Naik</span>, Georgios Sakkas, Saikat Chakraborty, Shuvendu Lahiri
</div>
<div class="publinks">
IEEE Transactions on Software Engineering, vol. 50, no. 9, pp. 2254-2268, Sept. 2024
[
<a href="https://ieeexplore.ieee.org/document/10606356" target="_blank">Paper</a>
]
</div>
</div>
</div>
<br>
<div class="row">
<div class="col-md-12 pubdiv" >
<h4>Workshop Papers</h4>
</div>
<div class="row">
<div class="col-md-12 pubdiv">
<h5 class="pubtitle">Learning to Walk over Relational Graphs of Source Code</h5>
<div class="pubauths">
Pardis Pashakhanloo, <span style="font-weight: bold">Aaditya Naik</span>, Hanjun Dai, Petros Maniatis, Mayur Naik
</div>
<div class="publinks">
Deep Learning for Code (<a href="https://dl4c.github.io/" target="_blank">DL4C</a>) Workshop @ <a href="https://iclr.cc/" target="_blank" style="font-weight: bold;">ICLR 2022</a>
[
<a href="https://openreview.net/pdf?id=SubGAoOWJWc" target="_blank">Paper</a>
],
[
<a href="https://openreview.net/forum?id=SubGAoOWJWc" target="_blank">Open Review Forum</a>
],
[
<a href="https://github.com/ppashakhanloo/CodeTrek" target="_blank">Code</a>
]
</div>
</div>
</div>
</div>