Merlin: specification inference for explicit information flow problems

By Benjamin Livshits, Aditya V. Nori, Sriram K. Rajamani, Anindya Banerjee. 2009

167 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems, Program Analysis & Verification, Privacy & Security, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Program Analysis & Verification

In Privacy & Security

In Signal Processing

Citations in the corpus

A user-guided approach to program analysis (2015, methodology): “In particular, specification inference techniques based on probabilistic inference [4, 17, 24] can be formulated as probabilistic analyses (as defined in § 3).”
GenProg: A Generic Method for Automatic Software Repair (2012, background): “, [55], [56]), formal specifications are rarely available; none of the programs presented in this paper are specified.”
Statistical Deobfuscation of Android Applications (2016, methodology): “Several works [22, 21, 16, 24] use graphical models to discover properties about programs such as function specifications and invariants.”
A Machine-learning Approach for Classifying and Categorizing Android Sources and Sinks (2014, methodology): “SUSI in comparison to Merlin does not need any information about the client program or application.”
Information Flow Analysis of Android Applications in DroidSafe (2015, background): “Merlin [45] is a probabilistic approach that employs a potentially incomplete list of sources, sinks, and sanitizers to calculate a more comprehensive list.”
Python probabilistic type inference with natural language support (2016, methodology): “Probabilistic inference has been successfully applied to many areas such as debugging and artificial intelligence [15, 11, 16, 12].”
Finding Likely Errors with Bayesian Specifications (2017, methodology): “MERLIN (Livshits et al. 2009) starts with a given initial specification and refines it through factor graph construction and inference.”
Bayesian specification learning for finding API usage errors (2017, background): “There is also a literature [7, 12, 27, 33, 36, 43, 47] on learning probabilistic specifications from programs.”
On the naturalness of software (2012, background): “This admittedly is a leap of faith; however, if it holds up (and we have found anecdotal evidence that it does, and some prior research implicitly makes a version of this assumption [36]) one can leverage this notion to build simple, scalable, and effective approximations in a wide variety of settin”