arXiv subjects

Shih-Chieh Su

Publications and source records attributed to Shih-Chieh Su.

10 recordsLinked to original sources

arXiv2025

AttnMod: Attention-Based New Art Styles

We introduce AttnMod, a training-free technique that modulates cross-attention in pre-trained diffusion models to generate novel, unpromptable art styles. The method is inspired by how a human artist might reinterpret a generated image, for example by emphasizing certain features, dispersing color, twisting silhouettes, or materializing unseen elements. AttnMod simulates this intent by altering how the text prompt conditions the image through attention during denoising. These targeted modulations enable diverse stylistic transformations without changing the prompt or retraining the model, and they expand the expressive capacity of text-to-image generation.

cs.CV↗

arXiv2023

Latent Painter

Latent diffusers revolutionized the generative AI and inspired creative art. When denoising the latent, the predicted original image at each step collectively animates the formation. However, the animation is limited by the denoising nature of the diffuser, and only renders a sharpening process. This work presents Latent Painter, which uses the latent as the canvas, and the diffuser predictions as the plan, to generate painting animation. Latent Painter also transits one generated image to another, which can happen between images from two different sets of checkpoints.

cs.CV↗

arXiv2022

BURSTT: Bustling Universe Radio Survey Telescope in Taiwan

Fast Radio Bursts (FRBs) are bright millisecond-duration radio transients that appear about 1,000 times per day, all-sky, for a fluence threshold 5 Jy ms at 600 MHz. The FRB radio-emission physics and the compact objects involved in these events are subjects of intense active debate. To better constrain source models, the Bustling Universe Radio Survey Telescope in Taiwan (BURSTT) is optimized to discover and localize a large sample of rare, high-fluence, nearby FRBs. This is the population most amenable to multi-messenger, multi-wavelength follow-up, allowing deeper understanding of source mechanisms. BURSTT will provide horizon-to-horizon sky coverage with a half power field-of-view (FoV) of $\sim$10$^{4}$ deg$^{2}$, a 400 MHz effective bandwidth between 300-800 MHz, and sub-arcsecond localization, made possible using outrigger stations hundreds to thousands of km from the main array. Initially, BURSTT will employ 256 antennas. After tests of various antenna designs and optimization of system performance we plan to expand to 2048 antennas. We estimate that BURSTT-256 will detect and localize $\sim$100 bright ($\geq$100 Jy ms) FRBs per year. Another advantage of BURSTT's large FoV and continuous operation will be greatly enhanced monitoring of FRBs for repetition. The current lack of sensitive all-sky observations likely means that many repeating FRBs are currently cataloged as single-event FRBs.

astro-ph.IM↗

arXiv2021

Frequency-induced Negative Magnetic Susceptibility in Epoxy/Magnetite Nanocomposites

The epoxy/magnetite nanocomposites express superparamagnetism under a static or low-frequency electromagnetic field. At the microwave frequency, said the X-band, the nanocomposites reveal an unexpected diamagnetism. To explain the intriguing phenomenon, we revisit the Debye relaxation law with the memory effect. The magnetization vector of the magnetite is unable to synchronize with the rapidly changing magnetic field, and it contributes to diamagnetism, a negative magnetic susceptibility for nanoparticles. The model just developed and the fitting result can not only be used to explain the experimental data in the X-band but also can be used to estimate the transition frequency between paramagnetism and diamagnetism.

cond-mat.mes-hall↗

arXiv2019

Channel Decomposition into Painting Actions

This work presents a method to decompose a convolutional layer of the deep neural network into painting actions. To behave like the human painter, these actions are driven by the cost simulating the hand movement, the paint color change, the stroke shape and the stroking style. To help planning, the Mask R-CNN is applied to detect the object areas and decide the painting order. The proposed painting system introduces a variety of extensions in artistic styles, based on the chosen parameters. Further experiments are performed to evaluate the channel penetration and the channel sensitivity on the strokes.

cs.GR↗

arXiv2017

Topical Behavior Prediction from Massive Logs

In this paper, we study the topical behavior in a large scale. We use the network logs where each entry contains the entity ID, the timestamp, and the meta data about the activity. Both the temporal and the spatial relationships of the behavior are explored with the deep learning architectures combing the recurrent neural network (RNN) and the convolutional neural network (CNN). To make the behavioral data appropriate for the spatial learning in the CNN, we propose several reduction steps to form the topical metrics and to place them homogeneously like pixels in the images. The experimental result shows both temporal and spatial gains when compared against a multilayer perceptron (MLP) network. A new learning framework called the spatially connected convolutional networks (SCCN) is introduced to predict the topical metrics more efficiently.

cs.LG↗

arXiv2017

Summarized Network Behavior Prediction

This work studies the entity-wise topical behavior from massive network logs. Both the temporal and the spatial relationships of the behavior are explored with the learning architectures combing the recurrent neural network (RNN) and the convolutional neural network (CNN). To make the behavioral data appropriate for the spatial learning in CNN, several reduction steps are taken to form the topical metrics and place them homogeneously like pixels in the images. The experimental result shows both the temporal- and the spatial- gains when compared to a multilayer perceptron (MLP) network. A new learning framework called spatially connected convolutional networks (SCCN) is introduced to more efficiently predict the behavior.

cs.LG↗

arXiv2016

Interacting with Massive Behavioral Data

In this short paper, we propose the split-diffuse (SD) algorithm that takes the output of an existing word embedding algorithm, and distributes the data points uniformly across the visualization space. The result improves the perceivability and the interactability by the human. We apply the SD algorithm to analyze the user behavior through access logs within the cyber security domain. The result, named the topic grids, is a set of grids on various topics generated from the logs. On the same set of grids, different behavioral metrics can be shown on different targets over different periods of time, to provide visualization and interaction to the human experts. Analysis, investigation, and other types of interaction can be performed on the topic grids more efficiently than on the output of existing dimension reduction methods. In addition to the cyber security domain, the topic grids can be further applied to other domains like e-commerce, credit card transaction, customer service to analyze the behavior in a large scale.

cs.LG↗

arXiv2016

Large Scale Behavioral Analytics via Topical Interaction

We propose the split-diffuse (SD) algorithm that takes the output of an existing dimension reduction algorithm, and distributes the data points uniformly across the visualization space. The result, called the topic grids, is a set of grids on various topics which are generated from the free-form text content of any domain of interest. The topic grids efficiently utilizes the visualization space to provide visual summaries for massive data. Topical analysis, comparison and interaction can be performed on the topic grids in a more perceivable way.

cs.LG↗

arXiv2016

Topic Grids for Homogeneous Data Visualization

We propose the topic grids to detect anomaly and analyze the behavior based on the access log content. Content-based behavioral risk is quantified in the high dimensional space where the topics are generated from the log. The topics are being projected homogeneously into a space that is perception- and interaction-friendly to the human experts.

cs.LG↗