Search

Multi-camera systems are indispensable in movies, TV shows, and other media. Selecting the appropriate camera at every timestamp has a decisive impact on production quality and audience preferences. Learning-based view recommendation frameworks can assist professionals in decision-making. However, they often struggle outside of their training domains. The scarcity of labeled multi-camera view recommendation datasets exacerbates the issue. Based on the insight that many videos are edited from the original multi-camera videos, we propose transforming regular videos into pseudo-labeled multi-camera view recommendation datasets. Promisingly, by training the model on pseudo-labeled datasets stemming from videos in the target domain, we achieve a 68% relative improvement in the model's accuracy in the target domain and bridge the accuracy gap between in-domain and never-before-seen domains., Comment: Accepted to VCIP 2024. Project page: https://eric11220.github.io/publication/VCIP24/

Spiking Neural Networks (SNNs) are well-suited for processing event streams from Dynamic Visual Sensors (DVSs) due to their use of sparse spike-based coding and asynchronous event-driven computation. To extract features from DVS objects, SNNs commonly use event-driven convolution with fixed kernel parameters. These filters respond strongly to features in specific orientations while disregarding others, leading to incomplete feature extraction. To improve the current event-driven convolution feature extraction capability of SNNs, we propose a DVS object recognition model that utilizes a trainable event-driven convolution and a spiking attention mechanism. The trainable event-driven convolution is proposed in this paper to update its convolution kernel through gradient descent. This method can extract local features of the event stream more efficiently than traditional event-driven convolution. Furthermore, the spiking attention mechanism is used to extract global dependence features. The classification performances of our model are better than the baseline methods on two neuromorphic datasets including MNIST-DVS and the more complex CIFAR10-DVS. Moreover, our model showed good classification ability for short event streams. It was shown that our model can improve the performance of event-driven convolutional SNNs for DVS objects., Comment: 14 pages, 2 figures

Automated plot generation for games enhances the player's experience by providing rich and immersive narrative experience that adapts to the player's actions. Traditional approaches adopt a symbolic narrative planning method which limits the scale and complexity of the generated plot by requiring extensive knowledge engineering work. Recent advancements use Large Language Models (LLMs) to drive the behavior of virtual characters, allowing plots to emerge from interactions between characters and their environments. However, the emergent nature of such decentralized plot generation makes it difficult for authors to direct plot progression. We propose a novel plot creation workflow that mediates between a writer's authorial intent and the emergent behaviors from LLM-driven character simulation, through a novel authorial structure called "abstract acts". The writers define high-level plot outlines that are later transformed into concrete character action sequences via an LLM-based narrative planning process, based on the game world state. The process creates "living stories" that dynamically adapt to various game world states, resulting in narratives co-created by the author, character simulation, and player. We present StoryVerse as a proof-of-concept system to demonstrate this plot creation workflow. We showcase the versatility of our approach with examples in different stories and game environments., Comment: Proceedings of the 19th international conference on the foundations of digital games 2024

We introduce HyperCLOVA X, a family of large language models (LLMs) tailored to the Korean language and culture, along with competitive capabilities in English, math, and coding. HyperCLOVA X was trained on a balanced mix of Korean, English, and code data, followed by instruction-tuning with high-quality human-annotated datasets while abiding by strict safety guidelines reflecting our commitment to responsible AI. The model is evaluated across various benchmarks, including comprehensive reasoning, knowledge, commonsense, factuality, coding, math, chatting, instruction-following, and harmlessness, in both Korean and English. HyperCLOVA X exhibits strong reasoning capabilities in Korean backed by a deep understanding of the language and cultural nuances. Further analysis of the inherent bilingual nature and its extension to multilingualism highlights the model's cross-lingual proficiency and strong generalization ability to untargeted languages, including machine translation between several language pairs and cross-lingual inference tasks. We believe that HyperCLOVA X can provide helpful guidance for regions or countries in developing their sovereign LLMs., Comment: 44 pages; updated authors list and fixed author names

We propose a two-stage estimation procedure for a copula-based model with semi-competing risks data, where the non-terminal event is subject to dependent censoring by the terminal event, and both events are subject to independent censoring. With a copula-based model, the marginal survival functions of individual event times are specified by semiparametric transformation models, and the dependence between the bivariate event times is specified by a parametric copula function. For the estimation procedure, in the first stage, the parameters associated with the marginal of the terminal event are estimated using only the corresponding observed outcomes, and in the second stage, the marginal parameters for the non-terminal event time and the copula parameter are estimated together via maximizing a pseudo-likelihood function based on the joint distribution of the bivariate event times. We derived the asymptotic properties of the proposed estimator and provided an analytic variance estimator for inference. Through simulation studies, we showed that our approach leads to consistent estimates with less computational cost and more robustness than the one-stage procedure developed in Chen (2012), where all parameters were estimated simultaneously. In addition, our approach demonstrates more desirable finite-sample performances over another existing two-stage estimation method proposed in Zhu et al. (2021). An R package PMLE4SCR is developed to implement our proposed method., Comment: 23 pages, 1 figure

We present numerical results for three-dimensional (3D) solitons with symmetries of the semi-vortex (SV) and mixed-mode (MM) types, which can be created in spinor Bose-Einstein condensates of Rydberg atoms under the action of the spin-orbit coupling (SOC). By means of systematic numerical computations, we demonstrate that the interplay of SOC and long-range spherically symmetric Rydberg interactions stabilize the 3D solitons, improving their resistance to collapse. We find how the stability range depends on the strengths of the SOC and Rydberg interactions and the soft-core atomic radius., Comment: to be published in Scientific Reports

The long-standing Gaussian product inequality (GPI) conjecture states that, for any centered $\mathbb{R}^n$-valued Gaussian random vector $(X_1, \dots, X_n)$ and any positive reals $\alpha_1, \dots, \alpha_n$, ${\bf E}[\prod_{j=1}^{n}|X_j|^{\alpha_j}]\ge \prod_{j=1}^{n}{\bf E}[|X_j|^{\alpha_j}]$. In this paper, we present some related inequalities for centered $\mathbb{R}^n$-valued Gaussian random vector $(X_1, \dots, X_n)$ when $\{\alpha_1, \dots, \alpha_n\}$ contains both positive and negative numbers., Comment: 15 pages

Let $B(n,p)$ denote a binomial random variable with parameters $n$ and $p$. Chv\'{a}tal's theorem says that for any fixed $n\geq 2$, as $m$ ranges over $\{0,\ldots,n\}$, the probability $q_m:=P(B(n,m/n)\leq m)$ is the smallest when $m$ is closest to $\frac{2n}{3}$. Motivated by this theorem, in this paper we consider the infimum value of the probability $P(X\leq \kappa E[X])$, where $\kappa$ is a positive real number, and $X$ is a random variable whose distribution belongs to some infinitely divisible distributions including the inverse Gaussian, log-normal, Gumbel and logistic distributions., Comment: 10 pages

Deep learning video analytic systems process live video feeds from multiple cameras with computer vision models deployed on edge or cloud. To optimize utility for these systems, which usually corresponds to query accuracy, efficient bandwidth management for the cameras competing for the fluctuating network resources is crucial. We propose DeepStream, a bandwidth efficient multi-camera video streaming system for deep learning video analytics. DeepStream addresses the challenge of limited and fluctuating bandwidth resources by offering several tailored solutions. We design a novel Regions of Interest detection (ROIDet) algorithm which can run in real time on resource constraint devices, such as Raspberry Pis, to remove spatial redundancy in video frames and reduce the amount of data to be transmitted. We also propose a content-aware bandwidth optimization framework and an Elastic Transmission Mechanism that exploits correlations among video contents. We implement DeepStream on Raspberry Pis and a desktop computer. Evaluations on real-world datasets show that DeepStream's ROIDet algorithm saves up to 54\% bandwidth with less than 1\% accuracy drop. Additionally,DeepStream improves utility by up to 23\% compared to baselines under the same bandwidth conditions.

Magnetic resonance imaging (MRI) using hyperpolarized noble gases provides a way to visualize the structure and function of human lung, but the long imaging time limits its broad research and clinical applications. Deep learning has demonstrated great potential for accelerating MRI by reconstructing images from undersampled data. However, most existing deep conventional neural networks (CNN) directly apply square convolution to k-space data without considering the inherent properties of k-space sampling, limiting k-space learning efficiency and image reconstruction quality. In this work, we propose an encoding enhanced (EN2) complex CNN for highly undersampled pulmonary MRI reconstruction. EN2 employs convolution along either the frequency or phase-encoding direction, resembling the mechanisms of k-space sampling, to maximize the utilization of the encoding correlation and integrity within a row or column of k-space. We also employ complex convolution to learn rich representations from the complex k-space data. In addition, we develop a feature-strengthened modularized unit to further boost the reconstruction performance. Experiments demonstrate that our approach can accurately reconstruct hyperpolarized 129Xe and 1H lung MRI from 6-fold undersampled k-space data and provide lung function measurements with minimal biases compared with fully-sampled image. These results demonstrate the effectiveness of the proposed algorithmic components and indicate that the proposed approach could be used for accelerated pulmonary MRI in research and clinical lung disease patient care.

360-degree video has become increasingly popular in content consumption. However, finding the viewing direction for important content within each frame poses a significant challenge. Existing approaches rely on either viewer input or algorithmic determination to select the viewing direction, but neither mode consistently outperforms the other in terms of content-importance. In this paper, we propose 360TripleView, the first view management system for 360-degree video that automatically infers and utilizes the better view mode for each frame, ultimately providing viewers with higher content-importance views. Through extensive experiments and a user study, we demonstrate that 360TripleView achieves over 90\% accuracy in inferring the better mode and significantly enhances content-importance compared to existing methods.