arXiv subjects

Hang Song

Publications and source records attributed to Hang Song.

At least 37 records · Page 2Linked to original sources

arXiv2025

Mapple: A Domain-Specific Language for Mapping Distributed Programs

Optimizing parallel programs for distributed systems is a complex task, often requiring significant code modifications. Task-based programming systems improve modularity by separating performance decisions from application logic, but their mapping interfaces are low-level. We introduce Mapple, a high-level, declarative programming interface for mapping distributed applications. Mapple provides transformation primitives to resolve dimensionality mismatches between task and processor spaces, including a key primitive, decompose, that helps minimize communication volume. We implement Mapple on top of the Legion runtime by translating Mapple mappers into its low-level C++ interface. Across nine applications, including six matrix multiplication algorithms and three scientific computing workloads, Mapple reduces mapper code size by 14x and enables performance improvements of up to 1.34x over expert-written C++ mappers. In addition, the decompose primitive achieves up to 1.83x improvement over existing dimensionality-resolution heuristics.

cs.DC↗

arXiv2025

Extraction of Moment Closures for Strongly Non-Equilibrium Flows via Machine Learning

We introduce a machine learning framework for moment-equation modeling of rarefied gas flows, addressing strongly non-equilibrium conditions inaccessible to conventional computational fluid dynamics. Our approach utilizes high-order moments and collision integrals, highly sensitive to non-equilibrium effects, as key predictive variables. Training datasets are created from one-dimensional steady shock simulations, and a methodology of computing collision integrals is developed. By learning thermodynamically consistent closures directly from DSMC data, our R13-ML model, combined with a discontinuous Galerkin solver for the transfer equations of moments, preserves physical structure and accurately predicts normal shock structures and generalizes to hypersonic and some unsteady, one-dimensional wave scenarios. This work bridges machine learning with continuum mechanics, offering a road map for high-fidelity aerothermal predictions in next-generation supersonic vehicles.

physics.flu-dyn↗

arXiv2025

Task-Based Programming for Adaptive Mesh Refinement in Compressible Flow Simulations

High-order solvers for compressible flows are vital in scientific applications. Adaptive mesh refinement (AMR) is a key technique for reducing computational cost by concentrating resolution in regions of interest. In this work, we develop an AMR-based numerical solver using Regent, a high-level programming language for the Legion programming model. We address several challenges associated with implementing AMR in Regent. These include dynamic data structures for patch refinement/coarsening, mesh validity enforcement, and reducing task launch overhead via task fusion. Experimental results show that task fusion achieves 18x speedup, while automated GPU kernel generation via simple annotations yields 9.7x speedup for the targeted kernel. We demonstrate our approach through simulations of two canonical compressible flow problems governed by the Euler equations.

cs.DC↗

arXiv2025

Mixed-R1: Unified Reward Perspective For Reasoning Capability in Multimodal Large Language Models

Recent works on large language models (LLMs) have successfully demonstrated the emergence of reasoning capabilities via reinforcement learning (RL). Although recent efforts leverage group relative policy optimization (GRPO) for MLLMs post-training, they constantly explore one specific aspect, such as grounding tasks, math problems, or chart analysis. There are no works that can leverage multi-source MLLM tasks for stable reinforcement learning. In this work, we present a unified perspective to solve this problem. We present Mixed-R1, a unified yet straightforward framework that contains a mixed reward function design (Mixed-Reward) and a mixed post-training dataset (Mixed-45K). We first design a data engine to select high-quality examples to build the Mixed-45K post-training dataset. Then, we present a Mixed-Reward design, which contains various reward functions for various MLLM tasks. In particular, it has four different reward functions: matching reward for binary answer or multiple-choice problems, chart reward for chart-aware datasets, IoU reward for grounding problems, and open-ended reward for long-form text responses such as caption datasets. To handle the various long-form text content, we propose a new open-ended reward named Bidirectional Max-Average Similarity (BMAS) by leveraging tokenizer embedding matching between the generated response and the ground truth. Extensive experiments show the effectiveness of our proposed method on various MLLMs, including Qwen2.5-VL and Intern-VL on various sizes. Our dataset and model are available at https://github.com/xushilin1/mixed-r1.

cs.CL↗

arXiv2024

Advancing Fine-Grained Visual Understanding with Multi-Scale Alignment in Multi-Modal Models

Multi-modal large language models (MLLMs) have achieved remarkable success in fine-grained visual understanding across a range of tasks. However, they often encounter significant challenges due to inadequate alignment for fine-grained knowledge, which restricts their ability to accurately capture local details and attain a comprehensive global perception. While recent advancements have focused on aligning object expressions with grounding information, they typically lack explicit integration of object images, which contain affluent information beyond mere texts or coordinates. To bridge this gap, we introduce a novel fine-grained visual knowledge alignment method that effectively aligns and integrates multi-scale knowledge of objects, including texts, coordinates, and images. This innovative method is underpinned by our multi-scale fine-grained enhancement data synthesis pipeline, which provides over 300K essential training data to enhance alignment and improve overall performance. Furthermore, we present TinyGroundingGPT, a series of compact models optimized for high-level alignments. With a scale of approximately 3B parameters, TinyGroundingGPT achieves outstanding results in grounding tasks while delivering performance comparable to larger MLLMs in complex visual scenarios.

cs.CV↗

arXiv2024

QCRD: Quality-guided Contrastive Rationale Distillation for Large Language Models

The deployment of large language models (LLMs) faces considerable challenges concerning resource constraints and inference efficiency. Recent research has increasingly focused on smaller, task-specific models enhanced by distilling knowledge from LLMs. However, prior studies have often overlooked the diversity and quality of knowledge, especially the untapped potential of negative knowledge. Constructing effective negative knowledge remains severely understudied. In this paper, we introduce a novel framework called quality-guided contrastive rationale distillation aimed at enhancing reasoning capabilities through contrastive knowledge learning. For positive knowledge, we enrich its diversity through temperature sampling and employ self-consistency for further denoising and refinement. For negative knowledge, we propose an innovative self-adversarial approach that generates low-quality rationales by sampling previous iterations of smaller language models, embracing the idea that one can learn from one's own weaknesses. A contrastive loss is developed to distill both positive and negative knowledge into smaller language models, where an online-updating discriminator is integrated to assess qualities of rationales and assign them appropriate weights, optimizing the training process. Through extensive experiments across multiple reasoning tasks, we demonstrate that our method consistently outperforms existing distillation techniques, yielding higher-quality rationales.

cs.CL↗

arXiv2024

UnifiedMLLM: Enabling Unified Representation for Multi-modal Multi-tasks With Large Language Model

Significant advancements has recently been achieved in the field of multi-modal large language models (MLLMs), demonstrating their remarkable capabilities in understanding and reasoning across diverse tasks. However, these models are often trained for specific tasks and rely on task-specific input-output formats, limiting their applicability to a broader range of tasks. This raises a fundamental question: Can we develop a unified approach to represent and handle different multi-modal tasks to maximize the generalizability of MLLMs? In this paper, we propose UnifiedMLLM, a comprehensive model designed to represent various tasks using a unified representation. Our model exhibits strong capabilities in comprehending the implicit intent of user instructions and preforming reasoning. In addition to generating textual responses, our model also outputs task tokens and grounding tokens, serving as indicators of task types and task granularity. These outputs are subsequently routed through the task router and directed to specific expert models for task completion. To train our model, we construct a task-specific dataset and an 100k multi-task dataset encompassing complex scenarios. Employing a three-stage training strategy, we equip our model with robust reasoning and task processing capabilities while preserving its generalization capacity and knowledge reservoir. Extensive experiments showcase the impressive performance of our unified representation approach across various tasks, surpassing existing methodologies. Furthermore, our approach exhibits exceptional scalability and generality. Our code, model, and dataset will be available at \url{https://github.com/lzw-lzw/UnifiedMLLM}.

cs.CL↗

arXiv2024

Small Distance Increment Method for Measuring Complex Permittivity With mmWave Radar

Measuring the complex permittivity of material is essential in many scenarios such as quality check and component analysis. Generally, measurement methods for characterizing the material are based on the usage of vector network analyzer, which is large and not easy for on-site measurement, especially in high frequency range such as millimeter wave (mmWave). In addition, some measurement methods require the destruction of samples, which is not suitable for non-destructive inspection. In this work, a small distance increment (SDI) method is proposed to non-destructively measure the complex permittivity of material. In SDI, the transmitter and receiver are formed as the monostatic radar, which is facing towards the material under test (MUT). During the measurement, the distance between radar and MUT changes with small increments and the signals are recorded at each position. A mathematical model is formulated to depict the relationship among the complex permittivity, distance increment, and measured signals. By fitting the model, the complex permittivity of MUT is estimated. To implement and evaluate the proposed SDI method, a commercial off-the-shelf mmWave radar is utilized and the measurement system is developed. Then, the evaluation was carried out on the acrylic plate. With the proposed method, the estimated complex permittivity of acrylic plate shows good agreement with the literature values, demonstrating the efficacy of SDI method for characterizing the complex permittivity of material.

eess.SP↗

arXiv2024

GroundingGPT:Language Enhanced Multi-modal Grounding Model

Multi-modal large language models have demonstrated impressive performance across various tasks in different modalities. However, existing multi-modal models primarily emphasize capturing global information within each modality while neglecting the importance of perceiving local information across modalities. Consequently, these models lack the ability to effectively understand the fine-grained details of input data, limiting their performance in tasks that require a more nuanced understanding. To address this limitation, there is a compelling need to develop models that enable fine-grained understanding across multiple modalities, thereby enhancing their applicability to a wide range of tasks. In this paper, we propose GroundingGPT, a language enhanced multi-modal grounding model. Beyond capturing global information like other multi-modal models, our proposed model excels at tasks demanding a detailed understanding of local information within the input. It demonstrates precise identification and localization of specific regions in images or moments in videos. To achieve this objective, we design a diversified dataset construction pipeline, resulting in a multi-modal, multi-granularity dataset for model training. The code, dataset, and demo of our model can be found at https: //github.com/lzw-lzw/GroundingGPT.

cs.CV↗

arXiv2023

QuDASH: Quantum-inspired rate adaptation approach for DASH video streaming

Internet traffic is dramatically increasing with the development of network technologies and video streaming traffic accounts for large amount within the total traffic, which reveals the importance to guarantee the quality of content delivery service. Based on the network conditions, adaptive bitrate (ABR) control is utilized as a common technique which can choose the proper bitrate to ensure the video streaming quality. In this paper, new bitrate control method, QuDASH is proposed by taking advantage of the emerging quantum technology. In QuDASH, the adaptive control model is developed using the quadratic unconstrained binary optimization (QUBO), which aims at increasing the average bitrate and decreasing the video rebuffering events to maximize the user quality of experience (QoE). In order to formulate the video control model, first the QUBO terms of different factors are defined regarding video quality, bitrate change, and buffer condition. Then, all the individual QUBO terms are merged to generate an objective function. By minimizing the QUBO objective function, the bitrate choice is determined from the solution. The control model is solved by Digital Annealer, which is a quantum-inspired computing technology. The evaluation of the proposed method is carried out by simulation with the throughput traces obtained in real world under different scenarios and the comparison with other methods is conducted. Experiment results demonstrated that the proposed QuDASH method has better performance in terms of QoE compared with other advanced ABR methods. In 68.2% of the examined cases, QuDASH achieves the highest QoE results, which shows the superiority of the QuDASH over conventional methods.

cs.NI↗

arXiv2022

Pensieve 5G: Implementation of RL-based ABR Algorithm for UHD 4K/8K Content Delivery on Commercial 5G SA/NR-DC Network

While the rollout of the fifth-generation mobile network (5G) is underway across the globe with the intention to deliver 4K/8K UHD videos, Augmented Reality (AR), and Virtual Reality (VR) content to the mass amounts of users, the coverage and throughput are still one of the most significant issues, especially in the rural areas, where only 5G in the low-frequency band are being deployed. This called for a high-performance adaptive bitrate (ABR) algorithm that can maximize the user quality of experience given 5G network characteristics and data rate of UHD contents. Recently, many of the newly proposed ABR techniques were machine-learning based. Among that, Pensieve is one of the state-of-the-art techniques, which utilized reinforcement-learning to generate an ABR algorithm based on observation of past decision performance. By incorporating the context of the 5G network and UHD content, Pensieve has been optimized into Pensieve 5G. New QoE metrics that more accurately represent the QoE of UHD video streaming on the different types of devices were proposed and used to evaluate Pensieve 5G against other ABR techniques including the original Pensieve. The results from the simulation based on the real 5G Standalone (SA) network throughput shows that Pensieve 5G outperforms both conventional algorithms and Pensieve with the average QoE improvement of 8.8% and 14.2%, respectively. Additionally, Pensieve 5G also performed well on the commercial 5G NR-NR Dual Connectivity (NR-DC) Network, despite the training being done solely using the data from the 5G Standalone (SA) network.

cs.NI↗

arXiv2022

RSSI-CSI Measurement and Variation Mitigation with Commodity WiFi Device

Owing to the plentiful information released by the commodity devices, WiFi signals have been widely studied for various wireless sensing applications. In many works, both received signal strength indicator (RSSI) and the channel state information (CSI) are utilized as the key factors for precise sensing. However, the calculation and relationship between RSSI and CSI is not explained in detail. Furthermore, there are few works focusing on the measurement variation of the WiFi signal which impacts the sensing results. In this paper, the relationship between RSSI and CSI is studied in detail and the measurement variation of amplitude and phase information is investigated by extensive experiments. In the experiments, transmitter and receiver are directly connected by power divider and RF cables and the signal transmission is quantitatively controlled by RF attenuators. By changing the intensity of attenuation, the measurement of RSSI and CSI is carried out under different conditions. From the results, it is found that in order to get a reliable measurement of the signal amplitude and phase by commodity WiFi, the attenuation of the channels should not exceed 60 dB. Meanwhile, the difference between two channels should be lower than 10 dB. An active control mechanism is suggested to ensure the measurement stability. The findings and criteria of this work is promising to facilitate more precise sensing technologies with WiFi signal.

eess.SP↗

arXiv2021

Adaptive video transmission using QUBO method and Digital Annealer based on Ising machine

With the dramatically increasing video streaming in the total network traffic, it is critical to develop effective algorithms to promote the content delivery service of high quality. Adaptive bitrate (ABR) control is the most essential technique which determines the proper bitrate to be chosen based on network conditions, thus realize high-quality video streaming. In this paper, a novel ABR strategy is proposed based on Ising machine by using the quadratic unconstrained binary optimization (QUBO) method and Digital Annealer (DA) for the first time. The proposed method is evaluated by simulation with the real-world measured throughput, and compared with other state-of-the-art methods. Experiment results show that the proposed QUBO-based method can outperform the existing methods, which demonstrating the superior of the proposed QUBO-based method.

cs.NI↗

arXiv2021

Scalable Parallel Linear Solver for Compact Banded Systems on Heterogeneous Architectures

A scalable algorithm for solving compact banded linear systems on distributed memory architectures is presented. The proposed method factorizes the original system into two levels of memory hierarchies, and solves it using parallel cyclic reduction on both distributed and shared memory. This method has a lower communication footprint across distributed memory partitions compared to conventional algorithms involving data transpose or re-partitioning. The algorithm developed in this work is generalized to cyclic compact banded systems with flexible data decompositions. For cyclic compact banded systems, the method is a direct solver with a deterministic operation and communication counts depending on the matrix size, its bandwidth, and the partition strategy. The implementation and runtime configuration details are discussed for performance optimization. Scalability is demonstrated on the linear solver as well as on a representative fluid mechanics application problem, in which the dominant computational cost is solving the cyclic tridiagonal linear systems of compact numerical schemes on a 3D periodic domain. The algorithm is particularly useful for solving the linear systems arising from the application of compact finite difference operators to a wide range of partial differential equation problems, such as but not limited to the numerical simulations of compressible turbulent flows, aeroacoustics, elastic-plastic wave propagation, and electromagnetics. It alleviates obstacles to their use on modern high performance computing hardware, where memory and computational power are distributed across nodes with multi-threaded processing units.

cs.DC↗

arXiv2020

The value chain of Industrial IoT and its reference framework for digitalization

Nowadays, we are rapidly moving beyond bespoke detailed solutions tailored for very specific problems, and we already build upon reusable and more general purpose infrastructures and tools, referring to them as IoT, Industrial IoT/Industry 4.0[1-3], etc. These are what will be discussed in this paper. When Industrial IoT (IIoT) is concerned about, the enormous innovation potential of IoT technologies are not only in the production of physical devices, but also in all activities performed by manufacturing industries, both in the pre-production (ideation, design, prototyping) and in the post-production (sales, training, maintenance, recycling) phases . It is also known that IIoT acquire and analyze data from connected devices, Cyber-Physical Systems (CPS), locations and people (e.g. operator); along with its contemporary new terms, such as 5G, Edge computing, and other ICT technologies with their applications[4] . More or less it is drawn upon on its combination with relative monitoring devices and actuators from operational technology (OT). IIoT helps regulate and monitor industrial systems [2], and it integrates/re-organize production resources flexibly, enhanced OT capability in the smart value chains enabling distributed decision-making of production.

cs.CY↗

arXiv2020

WiEps: Measurement of Dielectric Property with Commodity WiFi Device -- An application to Ethanol/Water Mixture

WiFi signal has become accessible everywhere, providing high-speed data transmission experience. Besides the communication service, channel state information (CSI) of the WiFi signals is widely employed for numerous Internet of Things (IoT) applications. Recently, most of these applications are based on analysis of the microwave reflections caused by physical movement of the objective. In this paper, a novel contactless wireless sensing technique named WiEps is developed to measure the dielectric properties of the material, exploiting the transmission characteristics of the WiFi signals. In WiEps, the material under test is placed between the transmitter antenna and receiver antenna. A theoretical model is proposed to quantitatively describe the relationship between CSI data and dielectric properties of the material. During the experiment, the phase and amplitude of the transmitted WiFi signals are extracted from the measured CSI data. The parameters of the theoretical model are calculated using measured data from the known materials. Then, WiEps is utilized to estimate the dielectric properties of unknown materials. The proposed technique is first applied to the ethanol/water mixtures. Then, additional liquids are measured for further verification. The estimated permittivities and conductivities show good agreement with the actual values, with the average error of 4.0% and 8.9%, respectively, indicating the efficacy of WiEps. By measuring the dielectric property, this technique is promising to be applied to new IoT applications using ubiquitous WiFi signals, such as food engineering, material manufacturing process monitoring, and security check.

eess.SP↗

arXiv2019

Synchronous Maneuver Searching and Trajectory Planning for Autonomous Vehicles in Dynamic Traffic Environments

In the real-time decision-making and local planning process of autonomous vehicles in dynamic environments, the autonomous driving system may fail to find a reasonable policy or even gets trapped in some situation due to the complexity of global tasks and the incompatibility between upper-level maneuver decisions with the low-level lower level trajectory planning. To solve this problem, this paper presents a synchronous maneuver searching and trajectory planning (SMSTP) algorithm based on the topological concept of homotopy. Firstly, a set of alternative maneuvers with boundary limits are enumerated on a multi-lane road. Instead of sampling numerous paths in the whole spatio-temporal space, we, for the first time, propose using Trajectory Profiles (TPs) to quickly construct the topological maneuvers represented by different routes, and put forward a corridor generation algorithm based on graph-search. The bounded corridor further constrains the maneuver's space in the spatial space. A step-wise heuristic optimization algorithm is then proposed to synchronously generate a feasible trajectory for each maneuver. To achieve real-time performance, we initialize the states to be optimized with the boundary constraints of maneuvers, and we set some heuristic states as terminal targets in the quadratic cost function. The solution of a feasible trajectory is always guaranteed only if a specific maneuver is given. The simulation and realistic driving-test experiments verified that the proposed SMSTP algorithm has a short computation time which is less than 37ms, and the experimental results showed the validity and effectiveness of the SMSTP algorithm.

cs.RO↗

arXiv2017

Discover and Learn New Objects from Documentaries

Despite the remarkable progress in recent years, detecting objects in a new context remains a challenging task. Detectors learned from a public dataset can only work with a fixed list of categories, while training from scratch usually requires a large amount of training data with detailed annotations. This work aims to explore a novel approach -- learning object detectors from documentary films in a weakly supervised manner. This is inspired by the observation that documentaries often provide dedicated exposition of certain object categories, where visual presentations are aligned with subtitles. We believe that object detectors can be learned from such a rich source of information. Towards this goal, we develop a joint probabilistic framework, where individual pieces of information, including video frames and subtitles, are brought together via both visual and linguistic links. On top of this formulation, we further derive a weakly supervised learning algorithm, where object model learning and training set mining are unified in an optimization procedure. Experimental results on a real world dataset demonstrate that this is an effective approach to learning new object detectors.

cs.CV↗