no code implementations • 8 Apr 2024 • Tham Yik Foong, Heng Zhang, Mao Po Yuan, Danilo Vasconcellos Vargas
In this paper, we demonstrated how a neural network integrated with Xenovert achieved better results in 4 out of 5 shifted datasets, saving the hurdle of retraining a machine learning model.
1 code implementation • 22 Mar 2024 • Bingli Liao, Danilo Vasconcellos Vargas
Large Language Models (LLMs) have shown remarkable capabilities, but their reasoning abilities and underlying mechanisms remain poorly understood.
1 code implementation • 7 Feb 2024 • Shashank Kotyan, PoYuan Mao, Danilo Vasconcellos Vargas
Deep neural networks are exploited using natural adversarial samples, which have no impact on human perception but are misclassified.
Evolutionary Algorithms Misclassification Rate - Natural Adversarial Samples
1 code implementation • 7 Dec 2023 • Shashank Kotyan, Ueda Tatsuya, Danilo Vasconcellos Vargas
While these methods effectively capture the overall sample distribution in the entire learned latent space, they tend to distort the structure of sample distributions within specific classes in the subset of the latent space.
no code implementations • 24 Nov 2023 • Mao Po-Yuan, Shashank Kotyan, Tham Yik Foong, Danilo Vasconcellos Vargas
To understand the impact of the initial seed vector on generated samples, we propose a reliability evaluation framework that evaluates the generated samples of a diffusion model when the initial seed vector is subjected to various synthetic shifts.
no code implementations • 22 Nov 2023 • Tham Yik Foong, Shashank Kotyan, Po Yuan Mao, Danilo Vasconcellos Vargas
Recent advances in text-to-image generators have led to substantial capabilities in image generation.
no code implementations • 16 Nov 2023 • Shashank Kotyan, Danilo Vasconcellos Vargas
Through this contribution, the paper aims to foster a deeper understanding of neural network limitations and proposes a practical approach to enhance their resilience in the face of evolving and unpredictable conditions.
no code implementations • 14 Nov 2023 • Shashank Kotyan, Danilo Vasconcellos Vargas
Neural networks have revolutionized various domains, exhibiting remarkable accuracy in tasks like natural language processing and computer vision.
no code implementations • 1 Nov 2023 • Shashank Kotyan, Danilo Vasconcellos Vargas
In conclusion, this work contributes to the ongoing research on Vision Transformers by introducing Dynamic Scanning Augmentation as a technique for improving the accuracy and robustness of ViT.
no code implementations • 16 Oct 2023 • Heng Zhang, Danilo Vasconcellos Vargas
The main idea is to apply equal updates from negative and positive feedback loops by symmetrical activation.
no code implementations • 27 Jul 2023 • Heng Zhang, Danilo Vasconcellos Vargas
Reservoir computing (RC), first applied to temporal signal processing, is a recurrent neural network in which neurons are randomly connected.
no code implementations • 19 Jun 2023 • Tham Yik Foong, Danilo Vasconcellos Vargas
However, we find that a reservoir that develops oscillatory activity without any external excitation can mimic the behaviour of a simple CPG in biological systems.
no code implementations • 4 Feb 2023 • Danilo Vasconcellos Vargas, Tham Yik Foong, Heng Zhang
The hurdle is that general patterns are difficult to define in terms of dynamical equations and designing a system that could learn by reordering itself is still to be seen.
no code implementations • 10 Jun 2021 • Shashank Kotyan, Danilo Vasconcellos Vargas
We also analyse how different adversarial samples distort the attention of the neural network compared to original samples.
no code implementations • 2 Sep 2020 • Danilo Vasconcellos Vargas, Bingli Liao, Takahiro Kanzaki
Thus, $\varphi$DNNs reveal that input recreation has strong benefits for artificial neural networks similar to biological ones, shedding light into the importance of purposely corrupting the input as well as pioneering an area of perception models based on GANs and autoencoders for robust recognition in artificial intelligence.
1 code implementation • 14 Jun 2020 • Danilo Vasconcellos Vargas, Toshitake Asabuki
Here, we propose a continual generalization of the chunking problem (an unsupervised problem), encompassing fixed and probabilistic chunks, discovery of temporal and causal structures and their continual variations.
no code implementations • 25 Sep 2019 • Danilo Vasconcellos Vargas, Shashank Kotyan, Moe Matsuki
The main idea lies in the fact that some features are present on unknown classes and that unknown classes can be defined as a combination of previous learned features without representation bias (a bias towards representation that maps only current set of input-outputs and their boundary).
1 code implementation • 27 Jun 2019 • Shashank Kotyan, Danilo Vasconcellos Vargas
By creating a novel neural architecture search with options for dense layers to connect with convolution layers and vice-versa as well as the addition of concatenation layers in the search, we were able to evolve an architecture that is inherently accurate on adversarial samples.
1 code implementation • 15 Jun 2019 • Shashank Kotyan, Danilo Vasconcellos Vargas, Moe Matsuki
A crucial step to understanding the rationale for this lack of robustness is to assess the potential of the neural networks' representation to encode the existing features.
1 code implementation • 14 Jun 2019 • Shashank Kotyan, Danilo Vasconcellos Vargas
There exists a vast number of adversarial attacks and defences for machine learning algorithms of various types which makes assessing the robustness of algorithms a daunting task.
no code implementations • 26 Apr 2019 • Shashank Kotyan, Danilo Vasconcellos Vargas, Venkanna U
Intrinsically, driving is a Markov Decision Process which suits well the reinforcement learning paradigm.
no code implementations • 18 Apr 2019 • Vinicius V. Melo, Danilo Vasconcellos Vargas, Wolfgang Banzhaf
Lexicase selection achieves very good solution quality by introducing ordered test cases.
no code implementations • 6 Mar 2019 • Danilo Vasconcellos Vargas, Junichi Murata, Hirotaka Takano
In this paper, both problems are solved together without approximations or simplifications.
no code implementations • 8 Feb 2019 • Danilo Vasconcellos Vargas, Jiawei Su
Deep neural networks were shown to be vulnerable to single pixel modifications.
no code implementations • 22 Jan 2019 • Di Li, Danilo Vasconcellos Vargas, Sakurai Kouichi
Here, we go beyond attacks to investigate, for the first time, universal rules, i. e., rules that are sample agnostic and therefore could turn any text sample in an adversarial one.
1 code implementation • 6 Jan 2019 • Danilo Vasconcellos Vargas, Junichi Murata
The combination of Spectrum Diversity with a unified neuron representation enables the algorithm to either surpass or equal NeuroEvolution of Augmenting Topologies (NEAT) on all of the five classes of problems tested.
1 code implementation • 2 Jan 2019 • Danilo Vasconcellos Vargas, Junichi Murata, Hirotaka Takano, Alexandre Claudio Botazzo Delbem
Structured evolutionary algorithms have been investigated for some time.
no code implementations • 20 Nov 2018 • Danilo Vasconcellos Vargas, Hirotaka Takano, Junichi Murata
Here, a variation of the first algorithm is proposed which uses a parameterless self organizing map (SOM).
no code implementations • 20 Nov 2018 • Danilo Vasconcellos Vargas, Hirotaka Takano, Junichi Murata
Experiments are conducted with the contingency training applied to neural networks over traditional datasets as well as datasets with additional irrelevant variables.
no code implementations • 20 Nov 2018 • Danilo Vasconcellos Vargas, Hirotaka Takano, Junichi Murata
In fact, the proposed algorithm possesses a dynamical population structure that self-organizes itself to better project the input space into a map.
no code implementations • 19 Sep 2018 • Danilo Vasconcellos Vargas, Hirotaka Takano, Junichi Murata
Moreover, NOTC is compared with NeuroEvolution of Augmenting Topologies (NEAT) in these problems, revealing a trade-off between the approaches.
no code implementations • 19 Apr 2018 • Jiawei Su, Danilo Vasconcellos Vargas, Kouichi Sakurai
The attack only requires modifying 5 pixels with 20. 44, 14. 76 and 22. 98 pixel values distortion.
no code implementations • 11 Feb 2018 • Jiawei Su, Danilo Vasconcellos Vargas, Sanjiva Prasad, Daniele Sgandurra, Yaokai Feng, Kouichi Sakurai
The Internet of Things (IoT) is an extension of the traditional Internet, which allows a very large number of smart devices, such as home appliances, network cameras, sensors and controllers to connect to one another to share information and improve user experiences.
6 code implementations • 24 Oct 2017 • Jiawei Su, Danilo Vasconcellos Vargas, Sakurai Kouichi
The results show that 67. 97% of the natural images in Kaggle CIFAR-10 test dataset and 16. 04% of the ImageNet (ILSVRC 2012) test images can be perturbed to at least one target class by modifying just one pixel with 74. 03% and 22. 91% confidence on average.