arXiv subjects

Xiao Ma

Publications and source records attributed to Xiao Ma.

At least 91 records · Page 5Linked to original sources

arXiv2024

Towards Timely Video Analytics Services at the Network Edge

Real-time video analytics services aim to provide users with accurate recognition results timely. However, existing studies usually fall into the dilemma between reducing delay and improving accuracy. The edge computing scenario imposes strict transmission and computation resource constraints, making balancing these conflicting metrics under dynamic network conditions difficult. In this regard, we introduce the age of processed information (AoPI) concept, which quantifies the time elapsed since the generation of the latest accurately recognized frame. AoPI depicts the integrated impact of recognition accuracy, transmission, and computation efficiency. We derive closed-form expressions for AoPI under preemptive and non-preemptive computation scheduling policies w.r.t. the transmission/computation rate and recognition accuracy of video frames. We then investigate the joint problem of edge server selection, video configuration adaptation, and bandwidth/computation resource allocation to minimize the long-term average AoPI over all cameras. We propose an online method, i.e., Lyapunov-based block coordinate descent (LBCD), to solve the problem, which decouples the original problem into two subproblems to optimize the video configuration/resource allocation and edge server selection strategy separately. We prove that LBCD achieves asymptotically optimal performance. According to the testbed experiments and simulation results, LBCD reduces the average AoPI by up to 10.94X compared to state-of-the-art baselines.

cs.DC↗

arXiv2024

Safety Alignment Should Be Made More Than Just a Few Tokens Deep

The safety alignment of current Large Language Models (LLMs) is vulnerable. Relatively simple attacks, or even benign fine-tuning, can jailbreak aligned models. We argue that many of these vulnerabilities are related to a shared underlying issue: safety alignment can take shortcuts, wherein the alignment adapts a model's generative distribution primarily over only its very first few output tokens. We refer to this issue as shallow safety alignment. In this paper, we present case studies to explain why shallow safety alignment can exist and provide evidence that current aligned LLMs are subject to this issue. We also show how these findings help explain multiple recently discovered vulnerabilities in LLMs, including the susceptibility to adversarial suffix attacks, prefilling attacks, decoding parameter attacks, and fine-tuning attacks. Importantly, we discuss how this consolidated notion of shallow safety alignment sheds light on promising research directions for mitigating these vulnerabilities. For instance, we show that deepening the safety alignment beyond just the first few tokens can often meaningfully improve robustness against some common exploits. Finally, we design a regularized finetuning objective that makes the safety alignment more persistent against fine-tuning attacks by constraining updates on initial tokens. Overall, we advocate that future safety alignment should be made more than just a few tokens deep.

cs.CR↗

arXiv2024

Redundancy-aware Action Spaces for Robot Learning

Joint space and task space control are the two dominant action modes for controlling robot arms within the robot learning literature. Actions in joint space provide precise control over the robot's pose, but tend to suffer from inefficient training; actions in task space boast data-efficient training but sacrifice the ability to perform tasks in confined spaces due to limited control over the full joint configuration. This work analyses the criteria for designing action spaces for robot manipulation and introduces ER (End-effector Redundancy), a novel action space formulation that, by addressing the redundancies present in the manipulator, aims to combine the advantages of both joint and task spaces, offering fine-grained comprehensive control with overactuated robot arms whilst achieving highly efficient robot learning. We present two implementations of ER, ERAngle (ERA) and ERJoint (ERJ), and we show that ERJ in particular demonstrates superior performance across multiple settings, especially when precise control over the robot configuration is required. We validate our results both in simulated and real robotic environments.

cs.RO↗

arXiv2024

Hierarchical Action Recognition: A Contrastive Video-Language Approach with Hierarchical Interactions

Video recognition remains an open challenge, requiring the identification of diverse content categories within videos. Mainstream approaches often perform flat classification, overlooking the intrinsic hierarchical structure relating categories. To address this, we formalize the novel task of hierarchical video recognition, and propose a video-language learning framework tailored for hierarchical recognition. Specifically, our framework encodes dependencies between hierarchical category levels, and applies a top-down constraint to filter recognition predictions. We further construct a new fine-grained dataset based on medical assessments for rehabilitation of stroke patients, serving as a challenging benchmark for hierarchical recognition. Through extensive experiments, we demonstrate the efficacy of our approach for hierarchical recognition, significantly outperforming conventional methods, especially for fine-grained subcategories. The proposed framework paves the way for hierarchical modeling in video understanding tasks, moving beyond flat categorization.

cs.CV↗

arXiv2024

Memory-efficient High-resolution OCT Volume Synthesis with Cascaded Amortized Latent Diffusion Models

Optical coherence tomography (OCT) image analysis plays an important role in the field of ophthalmology. Current successful analysis models rely on available large datasets, which can be challenging to be obtained for certain tasks. The use of deep generative models to create realistic data emerges as a promising approach. However, due to limitations in hardware resources, it is still difficulty to synthesize high-resolution OCT volumes. In this paper, we introduce a cascaded amortized latent diffusion model (CA-LDM) that can synthesis high-resolution OCT volumes in a memory-efficient way. First, we propose non-holistic autoencoders to efficiently build a bidirectional mapping between high-resolution volume space and low-resolution latent space. In tandem with autoencoders, we propose cascaded diffusion processes to synthesize high-resolution OCT volumes with a global-to-local refinement process, amortizing the memory and computational demands. Experiments on a public high-resolution OCT dataset show that our synthetic data have realistic high-resolution and global features, surpassing the capabilities of existing methods. Moreover, performance gains on two down-stream fine-grained segmentation tasks demonstrate the benefit of the proposed method in training deep learning models for medical imaging tasks. The code is public available at: https://github.com/nicetomeetu21/CA-LDM.

eess.IV↗

arXiv2024

A Universal List Decoding Algorithm with Application to Decoding of Polar Codes

This paper is concerned with a guessing codeword decoding (GCD) of linear block codes. Compared with the guessing noise decoding (GND), which is only efficient for high-rate codes, the GCD is efficient for not only high-rate codes but also low-rate codes. We prove that the GCD typically requires a fewer number of queries than the GND. Compared with the ordered statistics decoding (OSD), the GCD does not require the online Gaussian elimination (GE). In addition to limiting the maximum number of searches, we suggest limiting the radius of searches in terms of soft weights or tolerated performance loss to further reduce the decoding complexity, resulting in the so-called truncated GCD. The performance gap between the truncated GCD and the optimal decoding can be upper bounded approximately by the saddlepoint approach or other numerical approaches. The derived upper bound captures the relationship between the performance and the decoding parameters, enabling us to balance the performance and the complexity by optimizing the decoding parameters of the truncated GCD. We also introduce a parallel implementation of the (truncated) GCD algorithm to reduce decoding latency without compromising performance. Another contribution of this paper is the application of the GCD to the polar codes. We propose a multiple-bit-wise decoding algorithm over a pruned tree for the polar codes, referred to as the successive-cancellation list (SCL) decoding algorithm by GCD. First, we present a strategy for pruning the conventional polar decoding tree based on the complexity analysis rather than the specific bit patterns. Then we apply the GCD algorithm in parallel aided by the early stopping criteria to the leaves of the pruned tree. Simulation results show that, without any performance loss as justified by analysis, the proposed decoding algorithm can significantly reduce the decoding latency of the polar codes.

cs.IT↗

arXiv2024

DiTMoS: Delving into Diverse Tiny-Model Selection on Microcontrollers

Enabling efficient and accurate deep neural network (DNN) inference on microcontrollers is non-trivial due to the constrained on-chip resources. Current methodologies primarily focus on compressing larger models yet at the expense of model accuracy. In this paper, we rethink the problem from the inverse perspective by constructing small/weak models directly and improving their accuracy. Thus, we introduce DiTMoS, a novel DNN training and inference framework with a selector-classifiers architecture, where the selector routes each input sample to the appropriate classifier for classification. DiTMoS is grounded on a key insight: a composition of weak models can exhibit high diversity and the union of them can significantly boost the accuracy upper bound. To approach the upper bound, DiTMoS introduces three strategies including diverse training data splitting to increase the classifiers' diversity, adversarial selector-classifiers training to ensure synergistic interactions thereby maximizing their complementarity, and heterogeneous feature aggregation to improve the capacity of classifiers. We further propose a network slicing technique to alleviate the extra memory overhead incurred by feature aggregation. We deploy DiTMoS on the Neucleo STM32F767ZI board and evaluate it based on three time-series datasets for human activity recognition, keywords spotting, and emotion recognition, respectively. The experiment results manifest that: (a) DiTMoS achieves up to 13.4% accuracy improvement compared to the best baseline; (b) network slicing almost completely eliminates the memory overhead incurred by feature aggregation with a marginal increase of latency.

cs.LG↗

arXiv2024

Hierarchical Diffusion Policy for Kinematics-Aware Multi-Task Robotic Manipulation

This paper introduces Hierarchical Diffusion Policy (HDP), a hierarchical agent for multi-task robotic manipulation. HDP factorises a manipulation policy into a hierarchical structure: a high-level task-planning agent which predicts a distant next-best end-effector pose (NBP), and a low-level goal-conditioned diffusion policy which generates optimal motion trajectories. The factorised policy representation allows HDP to tackle both long-horizon task planning while generating fine-grained low-level actions. To generate context-aware motion trajectories while satisfying robot kinematics constraints, we present a novel kinematics-aware goal-conditioned control agent, Robot Kinematics Diffuser (RK-Diffuser). Specifically, RK-Diffuser learns to generate both the end-effector pose and joint position trajectories, and distill the accurate but kinematics-unaware end-effector pose diffuser to the kinematics-aware but less accurate joint position diffuser via differentiable kinematics. Empirically, we show that HDP achieves a significantly higher success rate than the state-of-the-art methods in both simulation and real-world.

cs.RO↗

arXiv2024

Mutual Information Regularized Offline Reinforcement Learning

The major challenge of offline RL is the distribution shift that appears when out-of-distribution actions are queried, which makes the policy improvement direction biased by extrapolation errors. Most existing methods address this problem by penalizing the policy or value for deviating from the behavior policy during policy improvement or evaluation. In this work, we propose a novel MISA framework to approach offline RL from the perspective of Mutual Information between States and Actions in the dataset by directly constraining the policy improvement direction. MISA constructs lower bounds of mutual information parameterized by the policy and Q-values. We show that optimizing this lower bound is equivalent to maximizing the likelihood of a one-step improved policy on the offline dataset. Hence, we constrain the policy improvement direction to lie in the data manifold. The resulting algorithm simultaneously augments the policy evaluation and improvement by adding mutual information regularizations. MISA is a general framework that unifies conservative Q-learning (CQL) and behavior regularization methods (e.g., TD3+BC) as special cases. We introduce 3 different variants of MISA, and empirically demonstrate that tighter mutual information lower bound gives better offline RL performance. In addition, our extensive experiments show MISA significantly outperforms a wide range of baselines on various tasks of the D4RL benchmark,e.g., achieving 742.9 total points on gym-locomotion tasks. Our code is available at https://github.com/sail-sg/MISA.

cs.LG↗

arXiv2024

Coding Theorems for Repetition and Superposition Codes over Binary-Input Output-Symmetric Channels

This paper is concerned with a class of low density generator matrix codes (LDGM), called repetition and superposition (RaS) codes, which have been proved to be capacity-achieving over binary-input output-symmetric (BIOS) channels in terms of bit-error rate (BER). We prove with a recently proposed framework that the RaS codes are also capacity-achieving over BIOS channels in terms of frame-error rate (FER). With this new framework, the theorem for the RaS codes can be generalized to source coding and joint source and channel coding (JSCC). In particular, we prove with this framework that the corresponding low-density parity-check (LDPC) codes, as an enlarged ensemble of quasi-cyclic LDPC (QC-LDPC) codes, can also achieve the capacity. To further improve the iterative decoding performance, we consider the convolutional RaS (Conv-RaS) code ensemble and prove it to be capacity-achieving over BIOS channels in terms of the first error event probability. The construction of Conv-RaS codes is flexible with rate (defined as the ratio of the input length to the encoding output length) ranging from less than one (typically for channel codes) to greater than one (typically for source codes), which can be implemented as a universal JSCC scheme, as confirmed by simulations.

cs.IT↗

arXiv2024

A Random Coding Approach to Performance Analysis of the Ordered Statistic Decoding with Local Constraints

This paper is concerned with the ordered statistic decoding with local constraints (LC-OSD) of binary linear block codes, which is a near maximum-likelihood decoding algorithm. Compared with the conventional OSD, the LC-OSD significantly reduces both the maximum and the average number of searches. The former is achieved by performing the serial list Viterbi algorithm (SLVA) or a two-way flipping pattern tree (FPT) algorithm with local constraints on the test error patterns, while the latter is achieved by incorporating tailored early termination criteria. The main objective of this paper is to explore the relationship between the performance of the LC-OSD and decoding parameters, such as the constraint degree and the maximum list size. To this end, we approximate the local parity-check matrix as a totally random matrix and then estimate the performance of the LC-OSD by analyzing with a saddlepoint approach the performance of random codes over the channels associated with the most reliable bits (MRBs). The random coding approach enables us to derive an upper bound on the performance and predict the average rank of the transmitted codeword in the list delivered by the LC-OSD. This allows us to balance the constraint degree and the maximum list size for the average (or maximum) time complexity reduction. Simulation results show that the approximation by random coding approach is numerically effective and powerful. Simulation results also show that the RS codes decoded by the LC-OSD can approach the random coding union (RCU) bounds, verifying the efficiency and universality of the LC-OSD.

cs.IT↗

arXiv2024

Guessing What, Noise or Codeword?

In this paper, we distinguish two guessing algorithms for decoding binary linear codes. One is the guessing noise decoding (GND) algorithm, and the other is the guessing codeword decoding (GCD) algorithm. We prove that the GCD is a maximum likelihood (ML) decoding algorithm and that the GCD is more efficient than GND for most practical applications. We also introduce several variants of ordered statistic decoding (OSD) to trade off the complexity of the Gaussian elimination (GE) and that of the guessing, which may find applications in decoding short block codes in the high signal-to-noise ratio (SNR) region.

cs.IT↗

arXiv2024

Resource-efficient In-orbit Detection of Earth Objects

With the rapid proliferation of large Low Earth Orbit (LEO) satellite constellations, a huge amount of in-orbit data is generated and needs to be transmitted to the ground for processing. However, traditional LEO satellite constellations, which downlink raw data to the ground, are significantly restricted in transmission capability. Orbital edge computing (OEC), which exploits the computation capacities of LEO satellites and processes the raw data in orbit, is envisioned as a promising solution to relieve the downlink burden. Yet, with OEC, the bottleneck is shifted to the inelastic computation capacities. The computational bottleneck arises from two primary challenges that existing satellite systems have not adequately addressed: the inability to process all captured images and the limited energy supply available for satellite operations. In this work, we seek to fully exploit the scarce satellite computation and communication resources to achieve satellite-ground collaboration and present a satellite-ground collaborative system named TargetFuse for onboard object detection. TargetFuse incorporates a combination of techniques to minimize detection errors under energy and bandwidth constraints. Extensive experiments show that TargetFuse can reduce detection errors by 3.4 times on average, compared to onboard computing. TargetFuse achieves a 9.6 times improvement in bandwidth efficiency compared to the vanilla baseline under the limited bandwidth budget constraint.

cs.NI↗

arXiv2023

InsActor: Instruction-driven Physics-based Characters

Generating animation of physics-based characters with intuitive control has long been a desirable task with numerous applications. However, generating physically simulated animations that reflect high-level human instructions remains a difficult problem due to the complexity of physical environments and the richness of human language. In this paper, we present InsActor, a principled generative framework that leverages recent advancements in diffusion-based human motion models to produce instruction-driven animations of physics-based characters. Our framework empowers InsActor to capture complex relationships between high-level human instructions and character motions by employing diffusion policies for flexibly conditioned motion planning. To overcome invalid states and infeasible state transitions in planned motions, InsActor discovers low-level skills and maps plans to latent skill sequences in a compact latent space. Extensive experiments demonstrate that InsActor achieves state-of-the-art results on various tasks, including instruction-driven motion generation and instruction-driven waypoint heading. Notably, the ability of InsActor to generate physically simulated animations using high-level human instructions makes it a valuable tool, particularly in executing long-horizon tasks with a rich set of instructions.

cs.CV↗

arXiv2023

Collaborative Weakly Supervised Video Correlation Learning for Procedure-Aware Instructional Video Analysis

Video Correlation Learning (VCL), which aims to analyze the relationships between videos, has been widely studied and applied in various general video tasks. However, applying VCL to instructional videos is still quite challenging due to their intrinsic procedural temporal structure. Specifically, procedural knowledge is critical for accurate correlation analyses on instructional videos. Nevertheless, current procedure-learning methods heavily rely on step-level annotations, which are costly and not scalable. To address this problem, we introduce a weakly supervised framework called Collaborative Procedure Alignment (CPA) for procedure-aware correlation learning on instructional videos. Our framework comprises two core modules: collaborative step mining and frame-to-step alignment. The collaborative step mining module enables simultaneous and consistent step segmentation for paired videos, leveraging the semantic and temporal similarity between frames. Based on the identified steps, the frame-to-step alignment module performs alignment between the frames and steps across videos. The alignment result serves as a measurement of the correlation distance between two videos. We instantiate our framework in two distinct instructional video tasks: sequence verification and action quality assessment. Extensive experiments validate the effectiveness of our approach in providing accurate and interpretable correlation analyses for instructional videos.

cs.CV↗

arXiv2023

Adjustable Robust Transformer for High Myopia Screening in Optical Coherence Tomography

Myopia is a manifestation of visual impairment caused by an excessively elongated eyeball. Image data is critical material for studying high myopia and pathological myopia. Measurements of spherical equivalent and axial length are the gold standards for identifying high myopia, but the available image data for matching them is scarce. In addition, the criteria for defining high myopia vary from study to study, and therefore the inclusion of samples in automated screening efforts requires an appropriate assessment of interpretability. In this work, we propose a model called adjustable robust transformer (ARTran) for high myopia screening of optical coherence tomography (OCT) data. Based on vision transformer, we propose anisotropic patch embedding (APE) to capture more discriminative features of high myopia. To make the model effective under variable screening conditions, we propose an adjustable class embedding (ACE) to replace the fixed class token, which changes the output to adapt to different conditions. Considering the confusion of the data at high myopia and low myopia threshold, we introduce the label noise learning strategy and propose a shifted subspace transition matrix (SST) to enhance the robustness of the model. Besides, combining the two structures proposed above, the model can provide evidence for uncertainty evaluation. The experimental results demonstrate the effectiveness and reliability of the proposed method. Code is available at: https://github.com/maxiao0234/ARTran.

cs.CV↗

arXiv2023

Beyond ChatBots: ExploreLLM for Structured Thoughts and Personalized Model Responses

Large language model (LLM) powered chatbots are primarily text-based today, and impose a large interactional cognitive load, especially for exploratory or sensemaking tasks such as planning a trip or learning about a new city. Because the interaction is textual, users have little scaffolding in the way of structure, informational "scent", or ability to specify high-level preferences or goals. We introduce ExploreLLM that allows users to structure thoughts, help explore different options, navigate through the choices and recommendations, and to more easily steer models to generate more personalized responses. We conduct a user study and show that users find it helpful to use ExploreLLM for exploratory or planning tasks, because it provides a useful schema-like structure to the task, and guides users in planning. The study also suggests that users can more easily personalize responses with high-level preferences with ExploreLLM. Together, ExploreLLM points to a future where users interact with LLMs beyond the form of chatbots, and instead designed to support complex user tasks with a tighter integration between natural language and graphical user interfaces.

cs.HC↗

arXiv2023

Energy and Time-Aware Inference Offloading for DNN-based Applications in LEO Satellites

In recent years, Low Earth Orbit (LEO) satellites have witnessed rapid development, with inference based on Deep Neural Network (DNN) models emerging as the prevailing technology for remote sensing satellite image recognition. However, the substantial computation capability and energy demands of DNN models, coupled with the instability of the satellite-ground link, pose significant challenges, burdening satellites with limited power intake and hindering the timely completion of tasks. Existing approaches, such as transmitting all images to the ground for processing or executing DNN models on the satellite, is unable to effectively address this issue. By exploiting the internal hierarchical structure of DNNs and treating each layer as an independent subtask, we propose a satellite-ground collaborative computation partial offloading approach to address this challenge. We formulate the problem of minimizing the inference task execution time and onboard energy consumption through offloading as an integer linear programming (ILP) model. The complexity in solving the problem arises from the combinatorial explosion in the discrete solution space. To address this, we have designed an improved optimization algorithm based on branch and bound. Simulation results illustrate that, compared to the existing approaches, our algorithm improve the performance by 10%-18%

cs.DC↗