A5100456758- Advanced Malware Detection Techniques
- Research Data Management Practices
- Formal Methods in Verification
- Security and Verification in Computing
- Adversarial Robustness in Machine Learning
- Software Testing and Debugging Techniques
- Particle Detector Development and Performance
- Cryptographic Implementations and Security
- Neutrino Physics Research
- Radiation Detection and Scintillator Technologies
- Scientific Computing and Data Management
- Astrophysics and Cosmic Phenomena
- Ethics and Social Impacts of AI
- Semantic Web and Ontologies
- Statistical Methods and Inference
- Parallel Computing and Optimization Techniques
- Online Learning and Analytics
- Intelligent Tutoring Systems and Adaptive Learning
- Statistical Methods and Bayesian Inference
- Recommender Systems and Techniques
- Evolutionary Algorithms and Applications
- E-Government and Public Services
- Machine Learning and Data Classification
- Advanced Causal Inference Techniques
- Online and Blended Learning
Purdue University West Lafayette
2025
Beijing Information Science & Technology University
2023-2024
South Dakota School of Mines and Technology
2024
Institute of Semiconductors
2024
Chinese Academy of Sciences
2024
University of Southern California
2018-2023
Leighton (United States)
2023
Southern California University for Professional Studies
2018-2022
Carnegie Mellon University
2021-2022
Kunming University of Science and Technology
2021-2022
Abstract Drug response variations amongst different individuals/populations are influenced by several factors including allele frequency differences of single nucleotide polymorphisms (SNPs) that functionally affect drug-response genes. Here, we aim to identify drugs potentially exhibit population in using SNP data mining and analytics. Ninety-one pairwise-comparisons >22,000,000 SNPs from the 1000 Genomes Project, across 14 populations, were performed ‘population-differentiated’...
As deep neural networks are increasingly being deployed in practice, their efficiency has become an important issue. While there compression techniques for reducing the network's size, energy consumption and computational requirement, they only demonstrate empirically that is no loss of accuracy, but lack formal guarantees compressed network, e.g., presence adversarial examples. Existing verification such as Reluplex, ReluVal, DeepPoly provide guarantees, designed analyzing a single network...
We propose a method for conducting algebraic program analysis (APA) incrementally in response to changes of the under analysis. APA is paradigm that consists two distinct steps: computing path expression succinctly summarizes set paths interest, and interpreting using properly-defined semantic algebra obtain properties interest. In this context, goal an incremental algorithm reduce time by leveraging intermediate results computed before changes. have made main contributions. First, we data...
The code generation modules inside modern compilers, which use a limited number of CPU registers to store large program variables, may introduce side-channel leaks even in software equipped with state-of-the-art countermeasures. We propose analysis and transformation based method eliminate such leaks. Our has type-based technique for detecting leaks, leverages Datalog-based declarative domain-specific optimizations achieve high efficiency accuracy. It also mitigation the compiler's backend,...
Cyber-physical systems are often safety-critical in that violations of safety properties may lead to catastrophes. We propose a method enforce the with real-valued signals by synthesizing runtime enforcer called shield. Whenever system violates property, shield, composed system, makes correction instantaneously ensure no erroneous output is generated combined system. While techniques for Boolean shields well understood, they do not handle ubiquitous cyber-physical systems, meaning their...
As neural networks make their way into safety-critical systems, where misbehavior can lead to catastrophes, there is a growing interest in certifying the equivalence of two structurally similar - problem known as differential verification. For example, compression techniques are often used practice for deploying trained on computationally- and energy-constrained devices, which raises question how faithfully compressed network mimics original network. Unfortunately, existing methods either...
Data poisoning aims to compromise a machine learning based software component by contaminating its training set change prediction results for test inputs. Existing methods deciding data-poisoning robustness have either poor accuracy or long running time and, more importantly, they can only certify some of the truly-robust cases, but remain inconclusive when certification fails. In other words, cannot falsify truly-non-robust cases. To overcome this limitation, we propose systematic testing...
We propose a data-driven method for synthesizing static analyses to detect side-channel information leaks in cryptographic software. Compared the conventional way of manually crafting such analyzers, which can be tedious, error prone and suboptimal, our learning-based technique is not only automated but also provably sound. Our analyzer consists set type-inference rules learned from training data, i.e., example code snippets annotated with ground truth. Internally, we use syntax-guided...
Automated static dataflow analysis is an effective technique for detecting security critical issues like sensitive data leak, and vulnerability to injection attacks. Ensuring high precision recall requires that context, field object sensitive. However, it challenging attain scale large industrial code bases. Compositional style analyses in which individual software components are analyzed separately, independent from their usage contexts, compute reusable summaries of components. This...
We consider the problem of assessing changing performance levels individual students as they go through online courses. This student (SP) modeling is a critical step for building adaptive teaching systems. Specifically, we conduct study how to utilize various types and large amounts log data train accurate machine learning (ML) models that predict future students. first use four very sets made available recently from distinct intelligent tutoring Our results include new ML approach defines...
We propose a method for synthesizing invariants that can help verify relational properties over two programs or different executions of program. Applications such include verifying functional equivalence, non-interference security, and continuity properties. Our generates invariant candidates using syntax guided synthesis (SyGuS) then filters them an SMT-solver based verifier, to ensure they are both inductive sufficient the property at hand. To improve performance, we learning techniques:...
Quantum computing is an emerging field of research and technology that harnesses a science called quantum mechanics to create computers with revolutionary capabilities. Although existing are limited in size prone significant errors, future might be capable performing tasks were once considered unimaginable using even the world’s most powerful supercomputers. This means could revolutionize many important areas our lives! In this article, we will explore by first reviewing how current work....
The government has always emphasized the decisive role of market mechanisms in resource allocation, but this is least obvious water resources market.To achieve goal promoting improvement supply structure and using thus improving efficiency, it necessary to consider both efficiency same source among different users cost between sources for user.Through creative use fuzzy optimization theory solve constructed Bi-level programming model, paper takes data Qingdao from 2011 2020 as an example...