Shuochao Yao

Ahmet Soyyigit, Shuochao Yao, Heechul Yun

In this work, we present a novel scheduling framework enabling anytime perception for deep neural network (DNN) based 3D object detection pipelines. We focus on computationally expensive region proposal network (RPN) and per-category multi-head detector components, which are common in 3D object detection pipelines, and make them deadline-aware. We propose a scheduling algorithm, which intelligently selects the subset of the components to make effective time and accuracy trade-off on the fly. We minimize accuracy loss of skipping some of the neural network sub-components by projecting previously detected objects onto the current scene through estimations. We apply our approach to a state-of-art 3D object detection network, PointPillars, and evaluate its performance on Jetson Xavier AGX using nuScenes dataset. Compared to the baselines, our approach significantly improve the network's accuracy under various deadline constraints.

Ahmet Soyyigit, Shuochao Yao, Heechul Yun

This work addresses the challenge of adapting dynamic deadline requirements for LiDAR object detection deep neural networks (DNNs). The computing latency of object detection is critically important to ensure safe and efficient navigation. However, state-of-the-art LiDAR object detection DNNs often exhibit significant latency, hindering their real-time performance on resource-constrained edge platforms. Therefore, a tradeoff between detection accuracy and latency should be dynamically managed at runtime to achieve optimum results. In this paper, we introduce VALO (Versatile Anytime algorithm for LiDAR Object detection), a novel data-centric approach that enables anytime computing of 3D LiDAR object detection DNNs. VALO employs a deadline-aware scheduler to selectively process input regions, making execution time and accuracy tradeoffs without architectural modifications. Additionally, it leverages efficient forecasting of past detection results to mitigate possible loss of accuracy due to partial processing of input. Finally, it utilizes a novel input reduction technique within its detection heads to significantly accelerate execution without sacrificing accuracy. We implement VALO on state-of-the-art 3D LiDAR object detection networks, namely CenterPoint and VoxelNext, and demonstrate its dynamic adaptability to a wide range of time constraints while achieving higher accuracy than the prior state-of-the-art. Code is available athttps://github.com/CSL-KU/VALO}{github.com/CSL-KU/VALO.

Huajie Shao, Dachun Sun, Jiahao Wu et al.

GitHub has become a popular social application platform, where a large number of users post their open source projects. In particular, an increasing number of researchers release repositories of source code related to their research papers in order to attract more people to follow their work. Motivated by this trend, we describe a novel item-item cross-platform recommender system, $\textit{paper2repo}$, that recommends relevant repositories on GitHub that match a given paper in an academic search system such as Microsoft Academic. The key challenge is to identify the similarity between an input paper and its related repositories across the two platforms, $\textit{without the benefit of human labeling}$. Towards that end, paper2repo integrates text encoding and constrained graph convolutional networks (GCN) to automatically learn and map the embeddings of papers and repositories into the same space, where proximity offers the basis for recommendation. To make our method more practical in real life systems, labels used for model training are computed automatically from features of user actions on GitHub. In machine learning, such automatic labeling is often called {\em distant supervision\/}. To the authors' knowledge, this is the first distant-supervised cross-platform (paper to repository) matching system. We evaluate the performance of paper2repo on real-world data sets collected from GitHub and Microsoft Academic. Results demonstrate that it outperforms other state of the art recommendation methods.

Renyuan Liu, Yuyang Leng, Kaiyan Liu et al.

Recent advancements in on-device training for deep neural networks have underscored the critical need for efficient activation compression to overcome the memory constraints of mobile and edge devices. As activations dominate memory usage during training and are essential for gradient computation, compressing them without compromising accuracy remains a key research challenge. While existing methods for dynamic activation quantization promise theoretical memory savings, their practical deployment is impeded by system-level challenges such as computational overhead and memory fragmentation. To address these challenges, we introduce DAF, a Dynamic Activation Framework that enables scalable and efficient on-device training through system-level optimizations. DAF achieves both memory- and time-efficient dynamic quantization training by addressing key system bottlenecks. It develops hybrid reduction operations tailored to the memory hierarchies of mobile and edge SoCs, leverages collaborative CPU-GPU bit-packing for efficient dynamic quantization, and implements an importance-aware paging memory management scheme to reduce fragmentation and support dynamic memory adjustments. These optimizations collectively enable DAF to achieve substantial memory savings and speedup without compromising model training accuracy. Evaluations on various deep learning models across embedded and mobile platforms demonstrate up to a $22.9\times$ reduction in memory usage and a $3.2\times$ speedup, making DAF a scalable and practical solution for resource-constrained environments.

Junyi Chen, Chuheng Du, Renyuan Liu et al.

Real-time LLM interactions demand streamed token generations, where text tokens are progressively generated and delivered to users while balancing two objectives: responsiveness (i.e., low time-to-first-token) and steady generation (i.e.,required time-between-tokens). Standard LLM serving systems suffer from the inflexibility caused by non-preemptive request scheduling and reactive memory management, leading to poor resource utilization and low request processing parallelism under request bursts. Therefore, we present TokenFlow, a novel LLM serving system with enhanced text streaming performance via preemptive request scheduling and proactive key-value (KV) cache management. TokenFlow dynamically prioritizes requests based on real-time token buffer occupancy and token consumption rate, while actively transferring KV cache between GPU and CPU memory in the background and overlapping I/O with computation to minimize request preemption overhead. Extensive experiments on Llama3-8B and Qwen2.5-32B across multiple GPUs (RTX 4090, A6000, H200) demonstrate that TokenFlow achieves up to 82.5% higher effective throughput (accounting for actual user consumption) while reducing P99 TTFT by up to 80.2%, without degrading overall token throughput.

Hongpeng Guo, Shuochao Yao, Zhe Yang et al.

Video cameras are pervasively deployed in city scale for public good or community safety (i.e. traffic monitoring or suspected person tracking). However, analyzing large scale video feeds in real time is data intensive and poses severe challenges to network and computation systems today. We present CrossRoI, a resource-efficient system that enables real time video analytics at scale via harnessing the videos content associations and redundancy across a fleet of cameras. CrossRoI exploits the intrinsic physical correlations of cross-camera viewing fields to drastically reduce the communication and computation costs. CrossRoI removes the repentant appearances of same objects in multiple cameras without harming comprehensive coverage of the scene. CrossRoI operates in two phases - an offline phase to establish cross-camera correlations, and an efficient online phase for real time video inference. Experiments on real-world video feeds show that CrossRoI achieves 42% - 65% reduction for network overhead and 25% - 34% reduction for response delay in real time video analytics applications with more than 99% query accuracy, when compared to baseline methods. If integrated with SotA frame filtering systems, the performance gains of CrossRoI reach 50% - 80% (network overhead) and 33% - 61% (end-to-end delay).

Shuochao Yao, Yifan Hao, Yiran Zhao et al.

The paper presents an efficient real-time scheduling algorithm for intelligent real-time edge services, defined as those that perform machine intelligence tasks, such as voice recognition, LIDAR processing, or machine vision, on behalf of local embedded devices that are themselves unable to support extensive computations. The work contributes to a recent direction in real-time computing that develops scheduling algorithms for machine intelligence tasks with anytime prediction. We show that deep neural network workflows can be cast as imprecise computations, each with a mandatory part and (several) optional parts whose execution utility depends on input data. The goal of the real-time scheduler is to maximize the average accuracy of deep neural network outputs while meeting task deadlines, thanks to opportunistic shedding of the least necessary optional parts. The work is motivated by the proliferation of increasingly ubiquitous but resource-constrained embedded devices (for applications ranging from autonomous cars to the Internet of Things) and the desire to develop services that endow them with intelligence. Experiments on recent GPU hardware and a state of the art deep neural network for machine vision illustrate that our scheme can increase the overall accuracy by 10%-20% while incurring (nearly) no deadline misses.

Huajie Shao, Zhisheng Xiao, Shuochao Yao et al.

This paper reviews the novel concept of controllable variational autoencoder (ControlVAE), discusses its parameter tuning to meet application needs, derives its key analytic properties, and offers useful extensions and applications. ControlVAE is a new variational autoencoder (VAE) framework that combines the automatic control theory with the basic VAE to stabilize the KL-divergence of VAE models to a specified value. It leverages a non-linear PI controller, a variant of the proportional-integral-derivative (PID) control, to dynamically tune the weight of the KL-divergence term in the evidence lower bound (ELBO) using the output KL-divergence as feedback. This allows us to precisely control the KL-divergence to a desired value (set point), which is effective in avoiding posterior collapse and learning disentangled representations. In order to improve the ELBO over the regular VAE, we provide simplified theoretical analysis to inform setting the set point of KL-divergence for ControlVAE. We observe that compared to other methods that seek to balance the two terms in VAE's objective, ControlVAE leads to better learning dynamics. In particular, it can achieve a good trade-off between reconstruction quality and KL-divergence. We evaluate the proposed method on three tasks: image generation, language modeling and disentangled representation learning. The results show that ControlVAE can achieve much better reconstruction quality than the other methods for comparable disentanglement. On the language modeling task, ControlVAE can avoid posterior collapse (KL vanishing) and improve the diversity of generated text. Moreover, our method can change the optimization trajectory, improving the ELBO and the reconstruction quality for image generation.

Wanzheng Zhu, Chao Zhang, Shuochao Yao et al.

We study the problem of modeling human mobility from semantic trace data, wherein each GPS record in a trace is associated with a text message that describes the user's activity. Existing methods fall short in unveiling human movement regularities, because they either do not model the text data at all or suffer from text sparsity severely. We propose SHMM, a multi-modal spherical hidden Markov model for semantics-rich human mobility modeling. Under the hidden Markov assumption, SHMM models the generation process of a given trace by jointly considering the observed location, time, and text at each step of the trace. The distinguishing characteristic of SHMM is the text modeling part. We use fixed-size vector representations to encode the semantics of the text messages, and model the generation of the l2-normalized text embeddings on a unit sphere with the von Mises-Fisher (vMF) distribution. Compared with other alternatives like multi-variate Gaussian, our choice of the vMF distribution not only incurs much fewer parameters, but also better leverages the discriminative power of text embeddings in a directional metric space. The parameter inference for the vMF distribution is non-trivial since it involves functional inversion of ratios of Bessel functions. We theoretically prove that: 1) the classical Expectation-Maximization algorithm can work with vMF distributions; and 2) while closed-form solutions are hard to be obtained for the M-step, Newton's method is guaranteed to converge to the optimal solution with quadratic convergence rate. We have performed extensive experiments on both synthetic and real-life data. The results on synthetic data verify our theoretical analysis; while the results on real-life data demonstrate that SHMM learns meaningful semantics-rich mobility models, outperforms state-of-the-art mobility models for next location prediction, and incurs lower training cost.

Huajie Shao, Haohong Lin, Qinmin Yang et al.

This paper challenges the common assumption that the weight $β$, in $β$-VAE, should be larger than $1$ in order to effectively disentangle latent factors. We demonstrate that $β$-VAE, with $β< 1$, can not only attain good disentanglement but also significantly improve reconstruction accuracy via dynamic control. The paper removes the inherent trade-off between reconstruction accuracy and disentanglement for $β$-VAE. Existing methods, such as $β$-VAE and FactorVAE, assign a large weight to the KL-divergence term in the objective function, leading to high reconstruction errors for the sake of better disentanglement. To mitigate this problem, a ControlVAE has recently been developed that dynamically tunes the KL-divergence weight in an attempt to control the trade-off to more a favorable point. However, ControlVAE fails to eliminate the conflict between the need for a large $β$ (for disentanglement) and the need for a small $β$. Instead, we propose DynamicVAE that maintains a different $β$ at different stages of training, thereby decoupling disentanglement and reconstruction accuracy. In order to evolve the weight, $β$, along a trajectory that enables such decoupling, DynamicVAE leverages a modified incremental PI (proportional-integral) controller, and employs a moving average as well as a hybrid annealing method to evolve the value of KL-divergence smoothly in a tightly controlled fashion. We theoretically prove the stability of the proposed approach. Evaluation results on three benchmark datasets demonstrate that DynamicVAE significantly improves the reconstruction accuracy while achieving disentanglement comparable to the best of existing methods. The results verify that our method can separate disentangled representation learning and reconstruction, removing the inherent tension between the two.

Chaoqi Yang, Ruijie Wang, Shuochao Yao et al.

Previous hypergraph expansions are solely carried out on either vertex level or hyperedge level, thereby missing the symmetric nature of data co-occurrence, and resulting in information loss. To address the problem, this paper treats vertices and hyperedges equally and proposes a new hypergraph formulation named the \emph{line expansion (LE)} for hypergraphs learning. The new expansion bijectively induces a homogeneous structure from the hypergraph by treating vertex-hyperedge pairs as "line nodes". By reducing the hypergraph to a simple graph, the proposed \emph{line expansion} makes existing graph learning algorithms compatible with the higher-order structure and has been proven as a unifying framework for various hypergraph expansions. We evaluate the proposed line expansion on five hypergraph datasets, the results show that our method beats SOTA baselines by a significant margin.

Huajie Shao, Shuochao Yao, Dachun Sun et al.

Variational Autoencoders (VAE) and their variants have been widely used in a variety of applications, such as dialog generation, image generation and disentangled representation learning. However, the existing VAE models have some limitations in different applications. For example, a VAE easily suffers from KL vanishing in language modeling and low reconstruction quality for disentangling. To address these issues, we propose a novel controllable variational autoencoder framework, ControlVAE, that combines a controller, inspired by automatic control theory, with the basic VAE to improve the performance of resulting generative models. Specifically, we design a new non-linear PI controller, a variant of the proportional-integral-derivative (PID) control, to automatically tune the hyperparameter (weight) added in the VAE objective using the output KL-divergence as feedback during model training. The framework is evaluated using three applications; namely, language modeling, disentangled representation learning, and image generation. The results show that ControlVAE can achieve better disentangling and reconstruction quality than the existing methods. For language modelling, it not only averts the KL-vanishing, but also improves the diversity of generated text. Finally, we also demonstrate that ControlVAE improves the reconstruction quality of generated images compared to the original VAE.

Chaoqi Yang, Ruijie Wang, Shuochao Yao et al.

Oversmoothing has been assumed to be the major cause of performance drop in deep graph convolutional networks (GCNs). In this paper, we propose a new view that deep GCNs can actually learn to anti-oversmooth during training. This work interprets a standard GCN architecture as layerwise integration of a Multi-layer Perceptron (MLP) and graph regularization. We analyze and conclude that before training, the final representation of a deep GCN does over-smooth, however, it learns anti-oversmoothing during training. Based on the conclusion, the paper further designs a cheap but effective trick to improve GCN training. We verify our conclusions and evaluate the trick on three citation networks and further provide insights on neighborhood aggregation in GCNs.

Chaoqi Yang, Jinyang Li, Ruijie Wang et al.

Much work on social media opinion polarization focuses on a flat categorization of stances (or orthogonal beliefs) of different communities from media traces. We extend in this work in two important respects. First, we detect not only points of disagreement between communities, but also points of agreement. In other words, we estimate community beliefs in the presence of overlap. Second, in lieu of flat categorization, we consider hierarchical belief estimation, where communities might be hierarchically divided. For example, two opposing parties might disagree on core issues, but within a party, despite agreement on fundamentals, disagreement might occur on further details. We call the resulting combined problem a hierarchical overlapping belief estimation problem. To solve it, this paper develops a new class of unsupervised Non-negative Matrix Factorization (NMF) algorithms, we call Belief Structured Matrix Factorization (BSMF). Our proposed unsupervised algorithm captures both the latent belief intersections and dissimilarities, as well as a hierarchical structure. We discuss the properties of the algorithm and evaluate it on both synthetic and real-world datasets. In the synthetic dataset, our model reduces error by 40%. In real Twitter traces, it improves accuracy by around 10%. The model also achieves 96.08% self-consistency in a sanity check.

Shuochao Yao, Ailing Piao, Wenjun Jiang et al.

Recent advances in deep learning motivate the use of deep neural networks in Internet-of-Things (IoT) applications. These networks are modelled after signal processing in the human brain, thereby leading to significant advantages at perceptual tasks such as vision and speech recognition. IoT applications, however, often measure physical phenomena, where the underlying physics (such as inertia, wireless signal propagation, or the natural frequency of oscillation) are fundamentally a function of signal frequencies, offering better features in the frequency domain. This observation leads to a fundamental question: For IoT applications, can one develop a new brand of neural network structures that synthesize features inspired not only by the biology of human perception but also by the fundamental nature of physics? Hence, in this paper, instead of using conventional building blocks (e.g., convolutional and recurrent layers), we propose a new foundational neural network building block, the Short-Time Fourier Neural Network (STFNet). It integrates a widely-used time-frequency analysis method, the Short-Time Fourier Transform, into data processing to learn features directly in the frequency domain, where the physics of underlying phenomena leave better foot-prints. STFNets bring additional flexibility to time-frequency analysis by offering novel nonlinear learnable operations that are spectral-compatible. Moreover, STFNets show that transforming signals to a domain that is more connected to the underlying physics greatly simplifies the learning process. We demonstrate the effectiveness of STFNets with extensive experiments. STFNets significantly outperform the state-of-the-art deep learning models in all experiments. A STFNet, therefore, demonstrates superior capability as the fundamental building block of deep neural networks for IoT applications for various sensor inputs.

Shuochao Yao, Yiran Zhao, Huajie Shao et al.

Deep neural networks show great potential as solutions to many sensing application problems, but their excessive resource demand slows down execution time, pausing a serious impediment to deployment on low-end devices. To address this challenge, recent literature focused on compressing neural network size to improve performance. We show that changing neural network size does not proportionally affect performance attributes of interest, such as execution time. Rather, extreme run-time nonlinearities exist over the network configuration space. Hence, we propose a novel framework, called FastDeepIoT, that uncovers the non-linear relation between neural network structure and execution time, then exploits that understanding to find network configurations that significantly improve the trade-off between execution time and accuracy on mobile and embedded devices. FastDeepIoT makes two key contributions. First, FastDeepIoT automatically learns an accurate and highly interpretable execution time model for deep neural networks on the target device. This is done without prior knowledge of either the hardware specifications or the detailed implementation of the used deep learning library. Second, FastDeepIoT informs a compression algorithm how to minimize execution time on the profiled device without impacting accuracy. We evaluate FastDeepIoT using three different sensing-related tasks on two mobile devices: Nexus 5 and Galaxy Nexus. FastDeepIoT further reduces the neural network execution time by $48\%$ to $78\%$ and energy consumption by $37\%$ to $69\%$ compared with the state-of-the-art compression algorithms.

Shuochao Yao, Yiran Zhao, Huajie Shao et al.

Recent advances in deep learning have led various applications to unprecedented achievements, which could potentially bring higher intelligence to a broad spectrum of mobile and ubiquitous applications. Although existing studies have demonstrated the effectiveness and feasibility of running deep neural network inference operations on mobile and embedded devices, they overlooked the reliability of mobile computing models. Reliability measurements such as predictive uncertainty estimations are key factors for improving the decision accuracy and user experience. In this work, we propose RDeepSense, the first deep learning model that provides well-calibrated uncertainty estimations for resource-constrained mobile and embedded devices. RDeepSense enables the predictive uncertainty by adopting a tunable proper scoring rule as the training criterion and dropout as the implicit Bayesian approximation, which theoretically proves its correctness.To reduce the computational complexity, RDeepSense employs efficient dropout and predictive distribution estimation instead of model ensemble or sampling-based method for inference operations. We evaluate RDeepSense with four mobile sensing applications using Intel Edison devices. Results show that RDeepSense can reduce around 90% of the energy consumption while producing superior uncertainty estimations and preserving at least the same model accuracy compared with other state-of-the-art methods.

Shuochao Yao, Yiran Zhao, Aston Zhang et al.

Recent advances in deep learning motivate the use of deep neutral networks in sensing applications, but their excessive resource needs on constrained embedded devices remain an important impediment. A recently explored solution space lies in compressing (approximating or simplifying) deep neural networks in some manner before use on the device. We propose a new compression solution, called DeepIoT, that makes two key contributions in that space. First, unlike current solutions geared for compressing specific types of neural networks, DeepIoT presents a unified approach that compresses all commonly used deep learning structures for sensing applications, including fully-connected, convolutional, and recurrent neural networks, as well as their combinations. Second, unlike solutions that either sparsify weight matrices or assume linear structure within weight matrices, DeepIoT compresses neural network structures into smaller dense matrices by finding the minimum number of non-redundant hidden elements, such as filters and dimensions required by each layer, while keeping the performance of sensing applications the same. Importantly, it does so using an approach that obtains a global view of parameter redundancies, which is shown to produce superior compression. We conduct experiments with five different sensing-related tasks on Intel Edison devices. DeepIoT outperforms all compared baseline algorithms with respect to execution time and energy consumption by a significant margin. It reduces the size of deep neural networks by 90% to 98.9%. It is thus able to shorten execution time by 71.4% to 94.5%, and decrease energy consumption by 72.2% to 95.7%. These improvements are achieved without loss of accuracy. The results underscore the potential of DeepIoT for advancing the exploitation of deep neural networks on resource-constrained embedded devices.

Shuochao Yao, Shaohan Hu, Yiran Zhao et al.

Mobile sensing applications usually require time-series inputs from sensors. Some applications, such as tracking, can use sensed acceleration and rate of rotation to calculate displacement based on physical system models. Other applications, such as activity recognition, extract manually designed features from sensor inputs for classification. Such applications face two challenges. On one hand, on-device sensor measurements are noisy. For many mobile applications, it is hard to find a distribution that exactly describes the noise in practice. Unfortunately, calculating target quantities based on physical system and noise models is only as accurate as the noise assumptions. Similarly, in classification applications, although manually designed features have proven to be effective, it is not always straightforward to find the most robust features to accommodate diverse sensor noise patterns and user behaviors. To this end, we propose DeepSense, a deep learning framework that directly addresses the aforementioned noise and feature customization challenges in a unified manner. DeepSense integrates convolutional and recurrent neural networks to exploit local interactions among similar mobile sensors, merge local interactions of different sensory modalities into global interactions, and extract temporal relationships to model signal dynamics. DeepSense thus provides a general signal estimation and classification framework that accommodates a wide range of applications. We demonstrate the effectiveness of DeepSense using three representative and challenging tasks: car tracking with motion sensors, heterogeneous human activity recognition, and user identification with biometric motion analysis. DeepSense significantly outperforms the state-of-the-art methods for all three tasks. In addition, DeepSense is feasible to implement on smartphones due to its moderate energy consumption and low latency