More Like this

1. Greedy Attack and Gumbel Attack: Generating Adversarial Examples on Discrete Data

   Yang, P ; Chen, J. ; Hsieh, C.-J. ; Wang, J.-L. ; Jordan, M. ( March 2020 , Journal of machine learning research)

   We present a probabilistic framework for studying adversarial attacks on discrete data. Based on this framework, we derive a perturbation-based method, Greedy Attack, and a scalable learning-based method, Gumbel Attack, that illustrate various tradeoffs in the design of attacks. We demonstrate the effectiveness of these methods using both quantitative metrics and human evaluation on various stateof-the-art models for text classification, including a word-based CNN, a character-based CNN and an LSTM. As an example of our results, we show that the accuracy of character-based convolutional networks drops to the level of random selection by modifying only five characters through Greedy Attack. 

   more » « less
2. Can Host-Based SDNs Rival the Traffic Engineering Abilities of Switch-Based SDNs?

   Lei, Yunsen ; Lanson, Julian P. ; Kaldawy, Remy M. ; Estrada, Jeffrey ; Shue, Craig A. ( October 2020 , IEEE Network of the Future (NoF))

   The software-defined networking (SDN) paradigm offers significant flexibility for network operators. However, the SDN community has focused on switch-based implementations, which pose several challenges. First, some may require significant hardware costs to upgrade a network. Further, fine-grained flow control in a switch-based SDN results in well-known, fundamental scalability limitations. These challenges may limit the reach of SDN technologies. In this work, we explore the extent to which host-based SDN agents can achieve feature parity with switch-based SDNs. Prior work has shown the potential of host-based SDNs for security and access control. Our study finds that with appropriate preparation, a host-based agent offers the same capabilities of switch-based SDNs in the remaining key area of traffic engineering, even in a legacy managed-switch network. We find the approach offers comparable performance to switch-based SDNs while eliminating the flow table scalability and cost concerns of switch-based SDN deployments. 

   more » « less
3. Multi-level Scene Modeling and Matching for Smartphone-Based Indoor Localization

   https://doi.org/10.1109/ISMAR-Adjunct.2019.00-22  

   Chen, Lidong ; Zou, Yin ; Chang, Yaohua ; Liu, Jinyun ; Lin, Benjamin ; Zhu, Zhigang ( October 2019 , Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR)

   Accurate indoor positioning has attracted a lot of attention for a variety of indoor location-based applications, with the rapid development of mobile devices and their onboard sensors. A hybrid indoor localization method is proposed based on single off-the-shelf smartphone, which takes advantage of its various onboard sensors, including camera, gyroscope and accelerometer. The proposed approach integrates three components: visual-inertial odometry (VIO), point-based area mapping, and plane-based area mapping. A simplified RANSAC strategy is employed in plane matching for the sake of processing time. Since Apple's augmented reality platform ARKit has many powerful high-level APIs on world tracking, plane detection and 3D modeling, a practical smartphone app for indoor localization is developed on an iPhone that can run ARKit. Experimental results demonstrate that our plane-based method can achieve an accuracy of about 0.3 meter, which is based on a much more lightweight model, but achieves more accurate results than the point-based model by directly using ARKit's area mapping. The size of the plane-based model is less than 2KB for a closed-loop corridor area of about 45m*15m, comparing to about 10MB of the point-based model. 

   more » « less
4. Position: IntelliBlox: A Toolkit for Integrating Block-Based Programming into Game-Based Learning Environments

   https://doi.org/10.1109/BB48857.2019.8941222  

   Taylor, Sandra ; Min, Wookhee ; Mott, Bradford ; Emerson, Andrew ; Smith, Andy ; Wiebe, Eric ; Lester, James ( October 2019 , 2019 IEEE Blocks and Beyond Workshop (B&B))

   Block-based programming languages reduce the need to learn low-level programming syntax while enabling novice learners to focus on computational thinking skills. Game-based learning environments have been shown to create effective and engaging learning experiences for students in a broad range of educational domains. The fusion of block-based programming with game-based learning offers significant potential to motivate learners to develop computational thinking skills. A key challenge educational game developers face in creating rich, interactive learning experiences that integrate computational thinking activities is the lack of an embeddable block-based programming toolkit. Current block-based programming languages, such as Blockly and Scratch, cannot be easily embedded into industry-standard 3D game engines. This paper presents IntelliBlox, a Blockly-inspired toolkit for the Unity cross-platform game engine that enables learners to create block-based programs within immersive game-based learning environments. Our experience using IntelliBlox suggests that it is an effective toolkit for integrating block-based programming challenges into game-based learning environments. 

   more » « less
5. Tag-based regulation of modules in genetic programming improves context-dependent problem solving

   https://doi.org/10.1007/s10710-021-09406-8  

   Lalejini, Alexander ; Moreno, Matthew Andres ; Ofria, Charles ( September 2021 , Genetic programming and evolvable machines)

   We introduce and experimentally demonstrate the utility of tag-based genetic regulation, a new genetic programming (GP) technique that allows programs to dynamically adjust which code modules to express.Tags are evolvable labels that provide a flexible mechanism for referencing code modules. Tag-based genetic regulation extends existing tag-based naming schemes to allow programs to “promote” and “repress” code modules in order to alter expression patterns. This extension allows evolution to structure a program as a gene regulatory network where modules are regulated based on instruction executions. We demonstrate the functionality of tag-based regulation on a range of program synthesis problems. We find that tag-based regulation improves problem-solving performance on context-dependent problems; that is, problems where programs must adjust how they respond to current inputs based on prior inputs. Indeed, the system could not evolve solutions to some context-dependent problems until regulation was added. Our implementation of tag-based genetic regulation is not universally beneficial, however. We identify scenarios where the correct response to a particular input never changes, rendering tag-based regulation an unneeded functionality that can sometimes impede adaptive evolution. Tag-based genetic regulation broadens our repertoire of techniques for evolving more dynamic genetic programs and can easily be incorporated into existing tag-enabled GP systems. 

   more » « less