GPLAG: Detection of Software Plagiarism by Program Dependence Graph Analysis

1
GPLAG Detection of Software Plagiarism by
Program Dependence Graph Analysis

• Chao Liu, Chen Chen,
• Jiawei Han, Philip S. Yu
• University of Illinois at Urbana-Champaign
• IBM T.J. Waston Research Center
• Presented by Chao Liu

2
Motivations

• Blossom of open-source projects
• SourceForge.net 125,090 projects as July 2006
• Convenience for software plagiarism?
• You can always find something online
• Core-part plagiarism
• Ripping off GUIs and irrelevant parts
• (Illegally) reuse the implementations of
  core-algorithms
• Our goal
• Efficient detection of core-part plagiarism

3
Challenges

• Effectiveness
• Professional plagiarists
• Automated plagiarism
• Efficiency
• Only a small part of code is plagiarized, how to
  detect it efficiently?

4
Outline

• Plagiarism Disguises
• Review of Plagiarism Detection
• GPLAG PDG-based Plagiarism Detection
• Efficiency and Scalability
• Experiments
• Conclusions

5
Original Program
A procedure in a program, called join
01 static void 02 make_blank (struct line
blank, int count) 03 04 int i 05
unsigned char buffer 06 struct field
fields 07 blank-gtnfields count 08
blank-gtbuf.size blank-gtbuf.length count
1 09 blank-gtbuf.buffer (char) xmalloc
(blank-gtbuf.size) 10 buffer (unsigned char
) blank-gtbuf.buffer 11 blank-gtfields
fields (struct field ) xmalloc (sizeof
(struct field) count) 12 for (i 0 i lt
count i) 13 ... 14 15
6
Disguise 1 Format Alteration
Insert comments and blanks
01 static void 02 make_blank (struct line
blank, int count) 03 04 int i 05
unsigned char buffer 06 struct field
fields 07 blank-gtnfields count //
initialization 08 blank-gtbuf.size
blank-gtbuf.length count 1 09
blank-gtbuf.buffer (char) xmalloc
(blank-gtbuf.size) 10 buffer (unsigned char
) blank-gtbuf.buffer 11 blank-gtfields
fields (struct field ) xmalloc (sizeof
(struct field) count) 12 for (i 0 i
lt count i) 13 ... 14 15
7
Disguise 2 Identifier Renaming
Rename variables consistently
01 static void 02 fill_content (struct line
fill, int num) 03 04 int i 05 unsigned
char buffer 06 struct field fields 07
fill-gtnfields num // initialization 08
fill-gtbuf.size fill-gtbuf.length num 1 09
fill-gtbuf.buffer (char) xmalloc
(fill-gtbuf.size) 10 buffer (unsigned char
) fill-gtbuf.buffer 11 fill-gtfields fields
(struct field ) xmalloc (sizeof (struct
field) num) 12 for (i 0 i lt num
i) 13 ... 14 15
8
Disguise 3 Statement Reordering
Reorder non-dependent statements
01 static void 02 fill_content (struct line
fill, int num) 03 04 int i 05 unsigned
char buffer 06 struct field fields 11
fill-gtfields fields (struct field )
xmalloc (sizeof (struct field) num) 08
fill-gtbuf.size fill-gtbuf.length num 1 09
fill-gtbuf.buffer (char) xmalloc
(fill-gtbuf.size) 10 buffer (unsigned char
) fill-gtbuf.buffer 07 fill-gtnfields num
// initialization 12 for (i 0 i lt num
i) 13 ... 14 15
9
Disguise 4 Control Replacement
Use equivalent control structure

• 01 static void
• 02 fill_content (struct line fill, int num)
• 03
• 04 int i
• 05 unsigned char buffer
• 06 struct field fields
• 11 fill-gtfields fields
• (struct field ) xmalloc (sizeof (struct
  field) num)
• 08 fill-gtbuf.size fill-gtbuf.length num
  1
• 09 fill-gtbuf.buffer (char) xmalloc
  (fill-gtbuf.size)
• 10 buffer (unsigned char )
  fill-gtbuf.buffer
• 07 fill-gtnfields num // initialization
• i 0
• while (i lt num)
• ...
• 15 i
• 16
• 17

10
Disguise 5 Code Insertion
Insert immaterial code

• 01 static void
• 02 fill_content (struct line fill, int num)
• 03
• 04 int i
• 05 unsigned char buffer
• 06 struct field fields
• 11 fill-gtfields fields
• (struct field ) xmalloc (sizeof (struct
  field) num)
• 08 fill-gtbuf.size fill-gtbuf.length num
  1
• 09 fill-gtbuf.buffer (char) xmalloc
  (fill-gtbuf.size)
• 10 buffer (unsigned char )
  fill-gtbuf.buffer
• 07 fill-gtnfields num // initialization
• i 0
• while (i lt num)
• ... for (int j 0 j lt i j)
• 15 i
• 16
• 17

11
Fully Disguised
12
Outline

• Plagiarism Disguises
• Review of Plagiarism Detection
• GPLAG PDG-based Plagiarism Detection
• Efficiency and Scalability
• Experiments
• Conclusions

13
Review of Plagiarism Detection

• String-based Baker et al. 1995
• A program represented as a string
• Blanks and comments ignored.
• AST-based Baxter et al. 1998, Kontogiannis et
  al. 1995
• A program is represented as an Abstract Syntax
  Tree (AST)
• Fragile to statement reordering, control
  replacement and code insertion
• Token-based Kamiya et al. 2002, Prechelt et al.
  2002
• Variables of the same type are mapped to the same
  token
• A program is represented as a token string
• Fingerprint of token strings is used for
  robustness Schleimer et al. 2003
• Partially robust to statement reordering, control
  replacement and code insertion
• Representatives Moss and JPlag

14
Outline

• Plagiarism Disguises
• Review of Plagiarism Detection
• GPLAG PDG-based Plagiarism Detection
• Efficiency and Scalability
• Experiments
• Conclusions

15
Graphic representation of source code
int sum(int array, int count) int i, sum
sum 0 for(i 0 i lt count i) sum
add(sum, arrayi) return sum

int add(int a, int b) return a b
16
Graphic representation of source code

int sum(int array, int count) int i, sum
sum 0 for(i 0 i lt count i) sum
add(sum, arrayi) return sum
int add(int a, int b) return a b
17
Control Dependency
int sum(int array, int count) int i, sum
sum 0 for(i 0 i lt count i) sum
add(sum, arrayi) return sum
int add(int a, int b) return a b
18
Data Dependency
int sum(int array, int count) int i, sum
sum 0 for(i 0 i lt count i) sum
add(sum, arrayi) return sum
int add(int a, int b) return a b
19
Plagiarism Detectible?
20
Corresponding PDGs
PDG for the Original Code
PDG for the Plagiarized Code
21
PDG-based Plagiarism Detection

• A program is represented as a set of PDGs
• Let g be a PDG of Procedure P in the original
  program
• Let g be a PDG of Procedure P in the plagiarism
  suspect
• Subgraph isomorphism implies plagiarism
• If g is subgraph isomorphic to g, P is likely
  plagiarized from P
• ?-isomorphism Graph g is ?-isomorphic to g if
  there exists a subgraph s of g such that s is
  subgraph isomorphic to g, and s ? g.
• If g is ?isomorphic to g, the PDG pair (g, g)
  is regarded as a plagiarized PDG pair, and is
  then returned to human beings for examination.

22
Advantages

• Robust because it is hard to overhaul PDGs
• Dependencies encode program logic
• Incentive of plagiarism

23
Outline

• Plagiarism Disguises
• Review of Plagiarism Detection
• GPLAG PDG-based Plagiarism Detection
• Efficiency and Scalability
• Experiments
• Conclusions

24
Efficiency and Scalability

• Search space
• If the original program has n procedures and the
  plagiarism suspect has m procedures
• nm subgraph isomorphism testings
• Pruning search space
• Lossless filter
• Statistical lossy filter

25
Lossless filter

• Interestingness
• PDGs smaller than an interesting size K are
  excluded from both sides
• ?-isomorphism definition
• A PDG pair (g, g) is discarded if g lt?g.

26
Lossy Filter

• Observation
• If procedure P is plagiarized from procedure P,
  its PDG g should look similar to g.
• So discard those dissimilar PDG pairs
• Requirement
• This filter must be light-weighted

27
Vertex Histogram

• Represent PDG g by
• h(g) (n1, n2, , nk),
• where ni is the frequency of the ith kind of
  vertices.
• Similarly, represent PDG g by
• h(g) (m1, m2, , mk).
• Direct similarity measurement?
• How to define a proper similarity threshold?
• Is thus defined threshold program-independent?

28
Hypothesis Testing-based Approach

• Basic idea
• Estimate a k-dimensional multinomial distribution
  from h(g)
• Test whether h(g) is likely an observation from
  
• If it is, g looks similar to g, and an
  isomorphism testing is needed.
• Otherwise, (g, g) is discarded

29
Technical Details
30
Technical Details (contd)
31
Work-flow of GPLAG

• PDGs are generated with Codesurfer
• Isomorphism testing is implemented with VFLib.

32
Outline

• Plagiarism Disguises
• Review of Plagiarism Detection
• GPLAG PDG-based Plagiarism Detection
• Efficiency and Scalability
• Experiments
• Conclusions

33
Experiment Design

• Subject programs
• Effectiveness
• Filter efficiency
• Core-part plagiarism detection

34
Effectiveness

• 2-hour manual plagiarism, but can be automated?
• GPLAG detects all plagiarized PDG pairs within 1
  second
• PDG isomorphism also reveals what plagiarism
  disguises are applied

35
Efficiency

• Subject programs
• bc, less and tar.
• Exact copy as plagiarism.
• Lossless and lossy filter
• Pruning PDG-pairs.
• Implication to overall time cost.

36
Pruning Uninteresting PDG-pairs

• Lossless only
• Lossless and lossy

37
Implication to Overall Time Cost

• Time-out for subgraph isomorphism testing, time
  hogs.
• Lossless filter does not save much time.
• Lossy filter significantly reduces the time cost.
• Major time saving comes from the avoidance of
  time hogs.

38
Detection of Core-part Plagiarism

• Lower time cost with lossy filter.
• Lower false positives with lossy filter.

39
Outline

• Plagiarism Disguises
• Review of Plagiarism Detection
• GPLAG PDG-based Plagiarism Detection
• Efficiency and Scalability
• Experiments
• Conclusions

40
Conclusions

• We developed a new algorithm GPLAG for software
  plagiarism detection
• It is more effective to fight against
  professional plagiarists
• We developed a statistical lossy filter, which
  improves the efficiency of GPLAG
• We experimentally verified the effectiveness and
  efficiency of GPLAG

41
Q A
Thank You!
42
References

• 1 B. S. Baker. On finding duplication and near
  duplication in large software systems. In Proc.
  of 2nd Working Conf. on Reverse Engineering,
  1995.
• 2 I. D. Baxter, A. Yahin, L. Moura, M.
  SantAnna, and L. Bier. Clone detection using
  abstract syntax trees. In Proc. of Int. Conf. on
  Software Maintenance, 1998.
• 3 K. Kontogiannis, M. Galler, and R. DeMori.
  Detecting code similarity using patterns. In
  Working Notes of 3rd Workshop on AI and Software
  Engineering, 1995.
• 4 T. Kamiya, S. Kusumoto, and K. Inoue.
  CCFinder a multilinguistic token-based code
  clone detection system for large scale source
  code. IEEE Trans. Softw. Eng., 28(7), 2002.
• 5 L. Prechelt, G. Malpohl, and M. Philippsen.
  Finding plagiarisms among a set of programs with
  JPlag. J. of Universal Computer Science, 8(11),
  2002.
• 6 S. Schleimer, D. S. Wilkerson, and A. Aiken.
  Winnowing local algorithms for document
  fingerprinting. SIGMOD, 2003.
• 7 V. B. Livshits and T. Zimmermann. Dynamine
  Finding common error patterns by mining software
  revision histories. In Proc. of 13th Int. Symp.
  on the Foundations of Software Engineering, 2005.
• 8 C. Liu, X. Yan, and J. Han. Mining control
  flow abnormality for logic error isolation. In In
  Proc. 2006 SIAM Int. Conf. on Data Mining, 2006.
• 9 C. Liu, X. Yan, H. Yu, J. Han, and P. S. Yu.
  Mining behavior graphs for backtrace of
  noncrashing bugs. In SDM, 2005.