-
Notifications
You must be signed in to change notification settings - Fork 2
/
index.html
415 lines (388 loc) · 19.9 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
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
<!DOCTYPE html>
<html>
<head>
<!-- Basic -->
<meta charset="utf-8" />
<meta http-equiv="X-UA-Compatible" content="IE=edge" />
<!-- Mobile Metas -->
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no" />
<!-- Site Metas -->
<meta name="keywords" content="" />
<meta name="description" content="" />
<meta name="author" content="" />
<title>Security of CodeLMs</title>
<!-- slider stylesheet -->
<!-- slider stylesheet -->
<link rel="stylesheet" type="text/css" href="https://cdnjs.cloudflare.com/ajax/libs/OwlCarousel2/2.3.4/assets/owl.carousel.min.css" />
<!-- bootstrap core css -->
<link rel="stylesheet" type="text/css" href="htmls/css/bootstrap.css" />
<!-- fonts style -->
<link href="https://fonts.googleapis.com/css?family=Open+Sans:400,700|Poppins:400,700&display=swap" rel="stylesheet">
<!-- Custom styles for this template -->
<link href="htmls/css/style.css" rel="stylesheet" />
<!-- responsive style -->
<link href="htmls/css/responsive.css" rel="stylesheet" />
</head>
<body>
<div class="hero_area">
<!-- header section strats -->
<header class="header_section">
<div class="container-fluid">
<!-- container-fluid-->
<nav class="navbar navbar-expand-lg custom_nav-container pt-3">
<button class="navbar-toggler" type="button" data-toggle="collapse" data-target="#navbarSupportedContent" aria-controls="navbarSupportedContent" aria-expanded="false" aria-label="Toggle navigation">
<span class="navbar-toggler-icon"></span>
</button>
<div class="collapse navbar-collapse" id="navbarSupportedContent">
<div class="d-flex ml-auto flex-column flex-lg-row align-items-center">
<ul class="navbar-nav">
<!-- <li class="nav-item active">
<a class="nav-link" href="index.html">Home <span class="sr-only">(current)</span></a>
</li> -->
<li class="nav-item">
<a class="nav-link" href="htmls/Adversarial Attacks.html"> Adversarial Attacks </a>
</li>
<li class="nav-item">
<a class="nav-link" href="htmls/Backdoor Attacks.html"> Backdoor Attacks </a>
</li>
<li class="nav-item">
<a class="nav-link" href="htmls/member.html"> Group Members </a>
</li>
</ul>
<!-- update -->
<!-- <div class="user_option">
<a href="htmls/member.html" style="color: #fff;">
<img src="htmls/images/user.png" alt="" style="height: 25px;">Group Members
</a>
</div> -->
</div>
</div>
</nav>
</div>
</header>
<!-- end header section -->
<!-- slider section -->
<section class=" slider_section position-relative">
<div class="container">
<div id="carouselExampleIndicators" class="carousel slide" data-ride="carousel">
<ol class="carousel-indicators">
<li data-target="#carouselExampleIndicators" data-slide-to="0" class="active"></li>
<li data-target="#carouselExampleIndicators" data-slide-to="1"></li>
<li data-target="#carouselExampleIndicators" data-slide-to="2"></li>
</ol>
<div class="carousel-inner">
<div class="carousel-item active">
<div class="row">
<div class="col">
<div class="detail-box">
<div>
<h2>
welcome to
</h2>
<h1>
Security of CodeLMs
</h1>
<p>
Language models for code (CodeLMs) have
emerged as powerful tools
for code-related tasks, outperforming traditional methods and standard
machine learning approaches.
</p>
<!-- <div class="">
<a href="">
Contact us
</a>
</div> -->
</div>
</div>
</div>
</div>
</div>
<div class="carousel-item">
<div class="row">
<div class="col">
<div class="detail-box">
<div>
<h2>
welcome to
</h2>
<h1>
Security of CodeLMs
</h1>
<p>
However, these models are susceptible to security vulnerabilities,
drawing increasing research attention from domains such as software engineering,
artificial intelligence, and cybersecurity.
Despite the growing body of research focused on the security of
CodeLMs, a comprehensive survey in this area remains absent.
</p>
<!-- <div class="">
<a href="">
Contact us
</a>
</div> -->
</div>
</div>
</div>
</div>
</div>
<div class="carousel-item">
<div class="row">
<div class="col">
<div class="detail-box">
<div>
<h2>
welcome to
</h2>
<h1>
Security of CodeLMs
</h1>
<p>
To address this gap, we systematically review relevant papers,
organizing them based on attack and defense strategies.
Furthermore, we provide an overview of commonly used language models, datasets,
and evaluation metrics, and highlight open-source
tools and promising directions for future research in securing CodeLMs.
</p>
<!-- <div class="">
<a href="">
Contact us
</a>
</div> -->
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- end slider section -->
</div>
<!-- do section -->
<section class="do_section layout_padding">
<div class="container">
<div class="heading_container">
<!-- <h2>
Paper Classification
</h2> -->
</div>
<div class="do_container">
<div class="box arrow-start arrow_bg">
<div class="img-box">
<img src="htmls/images/icons/adv.png" alt="">
</div>
<div class="detail-box">
<!-- <h6>
Paper Class One
</h6> -->
<a href="htmls/Adversarial Attacks.html" style="color: rgb(1, 1, 11); text-decoration: none; font-size: 16px; font-weight: bold;">
Adversarial Attacks
</a>
</div>
</div>
<div class="box arrow-middle arrow_bg">
<div class="img-box">
<img src="htmls/images/icons/backdoor.png" alt="">
</div>
<div class="detail-box">
<a href="htmls/Backdoor Attacks.html" style="color: rgb(1, 1, 26); text-decoration: none; font-size: 16px; font-weight: bold;">
Backdoor Attacks
</a>
</div>
</div>
<div class="box arrow-middle arrow_bg">
<div class="img-box">
<img src="htmls/images/d-3.png" alt="">
</div>
<div class="detail-box">
<a href="" style="color: rgb(1, 1, 14); text-decoration: none; font-size: 16px; font-weight: bold;">
Other Attacks
</a>
</div>
</div>
</div>
</div>
</section>
<!-- end do section -->
<!-- client section -->
<section class="client_section">
<div class="container">
<div class="heading_container">
<h2>
Publications
</h2>
</div>
<div class="carousel-wrap ">
<div class="owl-carousel">
<div class="item">
<div class="box">
<div class="img-box">
<a href="https://arxiv.org/abs/2305.17506"><img src="htmls/images/recent_papers/BadCode.png" alt=""></a>
</div>
<div class="detail-box">
<h5>
Backdooring Neural Code Search
</h5>
<!-- <img src="htmls/images/quote.png" alt=""> -->
<p>
Reusing off-the-shelf code snippets from online repositories is a common practice,
which significantly enhances the productivity of software developers.
To find desired code snippets, developers resort to code search engines through
natural language queries. Neural code search models are hence behind many such engines.
These models are based on deep learning and gain substantial attention due to their impressive
performance. However, the security aspect of these models is rarely studied.
Particularly, an adversary can inject a backdoor in neural code search models,
which return buggy or even vulnerable code with security/privacy issues.
This may impact the downstream software (e.g., stock trading systems and autonomous driving)
and cause financial loss and/or life-threatening incidents. In this paper,
we demonstrate such attacks are feasible and can be quite stealthy.
By simply modifying one variable/function name, the attacker can make
buggy/vulnerable code rank in the top 11%. Our attack BADCODE features a
special trigger generation and injection procedure, making the attack more
effective and stealthy. The evaluation is conducted on two neural code search
models and the results show our attack outperforms baselines by 60%. Our user
study demonstrates that our attack is more stealthy than the baseline by two times based on the F1 score.
</p>
</div>
</div>
</div>
<div class="item">
<div class="box">
<div class="img-box">
<a href="https://arxiv.org/abs/2410.02841"><img src="htmls/images/recent_papers/DICE.png" alt=""></a>
</div>
<div class="detail-box">
<h5>
Demonstration Attack against In-Context Learning for Code Intelligence
</h5>
<!-- <img src="htmls/images/quote.png" alt=""> -->
<p>
Recent advancements in large language models (LLMs) have revolutionized code intelligence
by improving programming productivity and alleviating challenges faced by software developers.
To further improve the performance of LLMs on specific code intelligence tasks and reduce training
costs, researchers reveal a new capability of LLMs: in-context learning (ICL). ICL allows LLMs to
learn from a few demonstrations within a specific context, achieving impressive results without
parameter updating. However, the rise of ICL introduces new security vulnerabilities in the
code intelligence field. In this paper, we explore a novel security scenario based on the
ICL paradigm, where attackers act as third-party ICL agencies and provide users with bad ICL
content to mislead LLMs outputs in code intelligence tasks. Our study demonstrates the feasibility
and risks of such a scenario, revealing how attackers can leverage malicious demonstrations to
construct bad ICL content and induce LLMs to produce incorrect outputs, posing significant threats
to system security. We propose a novel method to construct bad ICL content called DICE, which is
composed of two stages: Demonstration Selection and Bad ICL Construction, constructing targeted bad
ICL content based on the user query and transferable across different query inputs. Ultimately, our
findings emphasize the critical importance of securing ICL mechanisms to protect code intelligence
systems from adversarial manipulation.
</p>
</div>
</div>
</div>
<div class="item">
<div class="box">
<div class="img-box">
<a href="https://arxiv.org/abs/2408.04683"><img src="htmls/images/recent_papers/EliBadcode.png" alt=""></a>
</div>
<div class="detail-box">
<h5>
Eliminating Backdoors in Neural Code Models via Trigger Inversion
</h5>
<!-- <img src="htmls/images/quote.png" alt=""> -->
<p>
Neural code models (NCMs) have been widely used for addressing various code understanding tasks,
such as defect detection and clone detection. However, numerous recent studies reveal that such
models are vulnerable to backdoor attacks. Backdoored NCMs function normally on normal code snippets,
but exhibit adversary-expected behavior on poisoned code snippets injected with the adversary-crafted
trigger. It poses a significant security threat. For example, a backdoored defect detection model may
misclassify user-submitted defective code as non-defective. If this insecure code is then integrated
into critical systems, like autonomous driving systems, it could lead to life safety. However, there
is an urgent need for effective defenses against backdoor attacks targeting NCMs.
To address this issue, in this paper, we innovatively propose a backdoor defense technique based
on trigger inversion, called EliBadCode. EliBadCode first filters the model vocabulary for trigger
tokens to reduce the search space for trigger inversion, thereby enhancing the efficiency of the
trigger inversion. Then, EliBadCode introduces a sample-specific trigger position identification
method, which can reduce the interference of adversarial perturbations for subsequent trigger inversion,
thereby producing effective inverted triggers efficiently. Subsequently, EliBadCode employs a Greedy
Coordinate Gradient algorithm to optimize the inverted trigger and designs a trigger anchoring method
to purify the inverted trigger. Finally, EliBadCode eliminates backdoors through model unlearning.
We evaluate the effectiveness of EliBadCode in eliminating backdoor attacks against multiple NCMs used
for three safety-critical code understanding tasks. The results demonstrate that EliBadCode can
effectively eliminate backdoors while having minimal adverse effects on the normal functionality of the model.
</p>
</div>
</div>
</div>
<!-- <div class="item">
<div class="box">
<div class="img-box">
<a href="https://jos.org.cn/jos/article/abstract/nL021"><img src="htmls/images/recent_papers/EliBadcode.png" alt=""></a>
</div>
<div class="detail-box">
<h5>
深度代码模型安全综述
</h5>
<p>
With the significant success of deep learning technology in fields such as computer vision
and natural language processing, software engineering researchers have begun to explore its
integration into solving software engineering tasks. Existing research results indicate that
deep learning technology exhibits advantages in various software code-related tasks, such as
code retrieval and code summarization, that traditional methods and machine learning approaches
cannot match. These deep learning models, trained for code-related tasks, are collectively referred
to as deep code models. However, similar to natural language processing and image processing models,
deep code model security faces numerous challenges due to the vulnerability and lack of interpretability
of neural networks. It has become a focal point in the field of software engineering. In recent years,
researchers have proposed numerous attack and defense methods specific to deep code models.
Nevertheless, there is currently a lack of a systematic review of security research on deep code models,
hindering rapid understanding of the field for subsequent researchers. To address this gap and provide
a comprehensive overview of the current state, challenges, and latest research findings in this field,
this paper collected 32 relevant papers and categorized existing research results into two main classes:
backdoor attacks and defense techniques, and adversarial attacks and defense techniques.
The paper systematically organizes and summarizes the collected papers based on different
technological categories. Subsequently, the paper outlines commonly used experimental datasets
and evaluation metrics in this field. Finally, the paper analyzes key challenges faced by this
field and suggests feasible future research directions, aiming to provide valuable guidance for
further advancing the security of deep code.
</p>
</div>
</div>
</div> -->
</div>
</div>
</div>
</section>
<!-- footer section -->
<section class="container-fluid footer_section">
<p>
© 2024 All Rights Reserved By
<a href="https://html.design/">The Team</a>
</p>
</section>
<!-- footer section -->
<script type="text/javascript" src="htmls/js/jquery-3.4.1.min.js"></script>
<script type="text/javascript" src="htmls/js/bootstrap.js"></script>
<script type="text/javascript" src="https://cdnjs.cloudflare.com/ajax/libs/OwlCarousel2/2.3.4/owl.carousel.min.js">
</script>
<!-- owl carousel script
-->
<script type="text/javascript">
$(".owl-carousel").owlCarousel({
loop: true,
margin: 0,
navText: [],
center: true,
autoplay: true,
autoplayHoverPause: true,
responsive: {
0: {
items: 1
},
1000: {
items: 3
}
}
});
</script>
<!-- end owl carousel script -->
</body>
</html>