Yoonsung Kim

Jinwoo Hwang, Minsu Kim, Daeun Kim et al.

Modern retrospective analytics systems leverage cascade architecture to mitigate bottleneck for computing deep neural networks (DNNs). However, the existing cascades suffer two limitations: (1) decoding bottleneck is either neglected or circumvented, paying significant compute and storage cost for pre-processing; and (2) the systems are specialized for temporal queries and lack spatial query support. This paper presents CoVA, a novel cascade architecture that splits the cascade computation between compressed domain and pixel domain to address the decoding bottleneck, supporting both temporal and spatial queries. CoVA cascades analysis into three major stages where the first two stages are performed in compressed domain while the last one in pixel domain. First, CoVA detects occurrences of moving objects (called blobs) over a set of compressed frames (called tracks). Then, using the track results, CoVA prudently selects a minimal set of frames to obtain the label information and only decode them to compute the full DNNs, alleviating the decoding bottleneck. Lastly, CoVA associates tracks with labels to produce the final analysis results on which users can process both temporal and spatial queries. Our experiments demonstrate that CoVA offers 4.8x throughput improvement over modern cascade systems, while imposing modest accuracy loss.

Yoonsung Kim, Changhun Oh, Jinwoo Hwang et al.

Deep neural network (DNN) video analytics is crucial for autonomous systems such as self-driving vehicles, unmanned aerial vehicles (UAVs), and security robots. However, real-world deployment faces challenges due to their limited computational resources and battery power. To tackle these challenges, continuous learning exploits a lightweight "student" model at deployment (inference), leverages a larger "teacher" model for labeling sampled data (labeling), and continuously retrains the student model to adapt to changing scenarios (retraining). This paper highlights the limitations in state-of-the-art continuous learning systems: (1) they focus on computations for retraining, while overlooking the compute needs for inference and labeling, (2) they rely on power-hungry GPUs, unsuitable for battery-operated autonomous systems, and (3) they are located on a remote centralized server, intended for multi-tenant scenarios, again unsuitable for autonomous systems due to privacy, network availability, and latency concerns. We propose a hardware-algorithm co-designed solution for continuous learning, DaCapo, that enables autonomous systems to perform concurrent executions of inference, labeling, and training in a performant and energy-efficient manner. DaCapo comprises (1) a spatially-partitionable and precision-flexible accelerator enabling parallel execution of kernels on sub-accelerators at their respective precisions, and (2) a spatiotemporal resource allocation algorithm that strategically navigates the resource-accuracy tradeoff space, facilitating optimal decisions for resource allocation to achieve maximal accuracy. Our evaluation shows that DaCapo achieves 6.5% and 5.5% higher accuracy than a state-of-the-art GPU-based continuous learning systems, Ekya and EOMU, respectively, while consuming 254x less power.

Wonung Kim, Yubin Lee, Yoonsung Kim et al.

Transformers are the driving force behind today's Large Language Models (LLMs), serving as the foundation for their performance and versatility. Yet, their compute and memory costs grow with sequence length, posing scalability challenges for long-context inferencing. In response, the algorithm community is exploring alternative architectures, such as state space models (SSMs), linear attention, and recurrent neural networks (RNNs), which we refer to as post-transformers. This shift presents a key challenge: building a serving system that efficiently supports both transformer and post-transformer LLMs within a unified framework. To address this challenge, we analyze the performance characteristics of transformer and post-transformer LLMs. Despite their algorithmic differences, both are fundamentally limited by memory bandwidth under batched inference due to attention in transformers and state updates in post-transformers. Further analyses suggest two additional insights: (1) state update operations, unlike attention, incur high hardware cost, making per-bank PIM acceleration inefficient, and (2) different low-precision arithmetic methods offer varying accuracy-area tradeoffs, while we identify Microsoft's MX as the Pareto-optimal choice. Building on these insights, we design Pimba as an array of State-update Processing Units (SPUs), each shared between two banks to enable interleaved access to PIM. Each SPU includes a State-update Processing Engine (SPE) that comprises element-wise multipliers and adders using MX-based quantized arithmetic, enabling efficient execution of state update and attention operations. Our evaluation shows that, compared to LLM-optimized GPU and GPU+PIM systems, Pimba achieves up to 4.1x and 2.1x higher token generation throughput, respectively.

Junseok Park, Yoonsung Kim, Hee Bin Yoo et al.

Toddlers evolve from free exploration with sparse feedback to exploiting prior experiences for goal-directed learning with denser rewards. Drawing inspiration from this Toddler-Inspired Reward Transition, we set out to explore the implications of varying reward transitions when incorporated into Reinforcement Learning (RL) tasks. Central to our inquiry is the transition from sparse to potential-based dense rewards, which share optimal strategies regardless of reward changes. Through various experiments, including those in egocentric navigation and robotic arm manipulation tasks, we found that proper reward transitions significantly influence sample efficiency and success rates. Of particular note is the efficacy of the toddler-inspired Sparse-to-Dense (S2D) transition. Beyond these performance metrics, using Cross-Density Visualizer technique, we observed that transitions, especially the S2D, smooth the policy loss landscape, promoting wide minima that enhance generalization in RL models.

Changhun Oh, Seongryong Oh, Jinwoo Hwang et al.

3D Gaussian Splatting (3DGS) rendering in real-time on resource-constrained devices is essential for delivering immersive augmented and virtual reality (AR/VR) experiences. However, existing solutions struggle to achieve high frame rates, especially for high-resolution rendering. Our analysis identifies the sorting stage in the 3DGS rendering pipeline as the major bottleneck due to its high memory bandwidth demand. This paper presents Neo, which introduces a reuse-and-update sorting algorithm that exploits temporal redundancy in Gaussian ordering across consecutive frames, and devises a hardware accelerator optimized for this algorithm. By efficiently tracking and updating Gaussian depth ordering instead of re-sorting from scratch, Neo significantly reduces redundant computations and memory bandwidth pressure. Experimental results show that Neo achieves up to 10.0x and 5.6x higher throughput than state-of-the-art edge GPU and ASIC solution, respectively, while reducing DRAM traffic by 94.5% and 81.3%. These improvements make high-quality and low-latency on-device 3D rendering more practical.

Wongyu Kim, Hochang Lee, Sanghak Lee et al.

Query augmentation makes queries more meaningful by appending further information to the queries to find relevant documents. Current studies have proposed Large Language Model (LLM)-based embedders, which learn representation for embedding and generation for query augmentation in a multi-task manner by leveraging the generative capabilities of LLM. During inference, these jointly trained embedders have conducted query augmentation followed by embedding, showing effective results. However, augmenting every query leads to substantial embedding latency and query augmentation can be detrimental to performance for some queries. Also, previous methods have not been explored in multimodal environments. To tackle these problems, we propose M-Solomon, a universal multimodal embedder that can adaptively determine when to augment queries. Our approach first divides the queries of the training datasets into two groups at the dataset level. One includes queries that require augmentation and the other includes queries that do not. Then, we introduces a synthesis process that generates appropriate augmentations for queries that require them by leveraging a powerful Multimodal LLM (MLLM). Next, we present adaptive query augmentation. Through this step, M-Solomon can conduct query augmentation only when necessary by learning to generate synthetic augmentations with the prefix /augment for queries that demand them and to generate the simple string /embed for others. Experimental results showed that M-Solomon not only surpassed the baseline without augmentation by a large margin but also outperformed the baseline that always used augmentation, providing much faster embedding latency.

Jinwoo Hwang, Daeun Kim, Sangyeop Lee et al.

Recently, Video-Language Models (VideoLMs) have demonstrated remarkable capabilities, offering significant potential for flexible and powerful video query systems. These models typically rely on Vision Transformers (ViTs), which process video frames individually to extract visual embeddings. However, generating embeddings for large-scale videos requires ViT inferencing across numerous frames, posing a major hurdle to real-world deployment and necessitating solutions for integration into scalable video data management systems. This paper introduces Déjà Vu, a video-language query engine that accelerates ViT-based VideoLMs by reusing computations across consecutive frames. At its core is ReuseViT, a modified ViT model specifically designed for VideoLM tasks, which learns to detect inter-frame reuse opportunities, striking an effective balance between accuracy and reuse. Although ReuseViT significantly reduces computation, these savings do not directly translate into performance gains on GPUs. To overcome this, Déjà Vu integrates memory-compute joint compaction techniques that convert the FLOP savings into tangible performance gains. Evaluations on three VideoLM tasks show that Déjà Vu accelerates embedding generation by up to a 2.64x within a 2% error bound, dramatically enhancing the practicality of VideoLMs for large-scale video analytics.

Hyunseo Kim, Hyeonseo Yang, Taekyung Kim et al.

Active view selection in 3D scene reconstruction has been widely studied since training on informative views is critical for reconstruction. Recently, Neural Radiance Fields (NeRF) variants have shown promising results in active 3D reconstruction using uncertainty-guided view selection. They utilize uncertainties estimated with neural networks that encode scene geometry and appearance. However, the choice of uncertainty integration methods, either voxel-based or neural rendering, has conventionally depended on the types of scene uncertainty being estimated, whether geometric or appearance-related. In this paper, we introduce Colorized Surface Voxel (CSV)-based view selection, a new next-best view (NBV) selection method exploiting surface voxel-based measurement of uncertainty in scene appearance. CSV encapsulates the uncertainty of estimated scene appearance (e.g., color uncertainty) and estimated geometric information (e.g., surface). Using the geometry information, we interpret the uncertainty of scene appearance 3D-wise during the aggregation of the per-voxel uncertainty. Consequently, the uncertainty from occluded and complex regions is recognized under challenging scenarios with limited input data. Our method outperforms previous works on popular datasets, DTU and Blender, and our new dataset with imbalanced viewpoints, showing that the CSV-based view selection significantly improves performance by up to 30%.

Kibeom Kim, Min Whoo Lee, Yoonsung Kim et al.

Learning in a multi-target environment without prior knowledge about the targets requires a large amount of samples and makes generalization difficult. To solve this problem, it is important to be able to discriminate targets through semantic understanding. In this paper, we propose goal-aware cross-entropy (GACE) loss, that can be utilized in a self-supervised way using auto-labeled goal states alongside reinforcement learning. Based on the loss, we then devise goal-discriminative attention networks (GDAN) which utilize the goal-relevant information to focus on the given instruction. We evaluate the proposed methods on visual navigation and robot arm manipulation tasks with multi-target environments and show that GDAN outperforms the state-of-the-art methods in terms of task success ratio, sample efficiency, and generalization. Additionally, qualitative analyses demonstrate that our proposed method can help the agent become aware of and focus on the given instruction clearly, promoting goal-directed behavior.

Yeonwoo Jeong, Yoonsung Kim, Hyun Oh Song

We develop hierarchically quantized efficient embedding representations for similarity-based search and show that this representation provides not only the state of the art performance on the search accuracy but also provides several orders of speed up during inference. The idea is to hierarchically quantize the representation so that the quantization granularity is greatly increased while maintaining the accuracy and keeping the computational complexity low. We also show that the problem of finding the optimal sparse compound hash code respecting the hierarchical structure can be optimized in polynomial time via minimum cost flow in an equivalent flow network. This allows us to train the method end-to-end in a mini-batch stochastic gradient descent setting. Our experiments on Cifar100 and ImageNet datasets show the state of the art search accuracy while providing several orders of magnitude search speedup respectively over exhaustive linear search over the dataset.