arXiv subjects

Xinyi Zhang

Publications and source records attributed to Xinyi Zhang.

At least 109 records · Page 6Linked to original sources

arXiv2023

Learning Efficient Policies for Picking Entangled Wire Harnesses: An Approach to Industrial Bin Picking

Wire harnesses are essential connecting components in manufacturing industry but are challenging to be automated in industrial tasks such as bin picking. They are long, flexible and tend to get entangled when randomly placed in a bin. This makes it difficult for the robot to grasp a single one in dense clutter. Besides, training or collecting data in simulation is challenging due to the difficulties in modeling the combination of deformable and rigid components for wire harnesses. In this work, instead of directly lifting wire harnesses, we propose to grasp and extract the target following a circle-like trajectory until it is untangled. We learn a policy from real-world data that can infer grasps and separation actions from visual observation. Our policy enables the robot to efficiently pick and separate entangled wire harnesses by maximizing success rates and reducing execution time. To evaluate our policy, we present a set of real-world experiments on picking wire harnesses. Our policy achieves an overall 84.6% success rate compared with 49.2% in baseline. We also evaluate the effectiveness of our policy under different clutter scenarios using unseen types of wire harnesses. Results suggest that our approach is feasible for handling wire harnesses in industrial bin picking.

cs.RO↗

arXiv2022

Slow-roll inflation in $f(R,T)$ gravity with a $RT$ mixing term

We consider slow-roll inflationary models in a class of modified theories of gravity which contains non-minimal curvature-inflaton couplings, i.e., the $f(R,T)$ gravity, where $R$ is the Ricci scalar and $T$ is the trace of the inflaton energy-momentum tensor. On top of the minimally coupled $T$ that has been widely investigated in the literature, we further include a $RT$ mixing term in the theory. This mixing term introduces non-minimal derivative couplings and plays an important role in inflationary dynamics. Taking chaotic and natural inflation as examples, we find that the predictions for spectral tilt and the tensor-to-scalar ratio are sensitive to the existence of the $RT$ mixing term. In particular, by turning on this mixing term, it is possible to bring chaotic and natural inflation into better agreement with observational data.

gr-qc↗

arXiv2022

HySAGE: A Hybrid Static and Adaptive Graph Embedding Network for Context-Drifting Recommendations

The recent popularity of edge devices and Artificial Intelligent of Things (AIoT) has driven a new wave of contextual recommendations, such as location based Point of Interest (PoI) recommendations and computing resource-aware mobile app recommendations. In many such recommendation scenarios, contexts are drifting over time. For example, in a mobile game recommendation, contextual features like locations, battery, and storage levels of mobile devices are frequently drifting over time. However, most existing graph-based collaborative filtering methods are designed under the assumption of static features. Therefore, they would require frequent retraining and/or yield graphical models burgeoning in sizes, impeding their suitability for context-drifting recommendations. In this work, we propose a specifically tailor-made Hybrid Static and Adaptive Graph Embedding (HySAGE) network for context-drifting recommendations. Our key idea is to disentangle the relatively static user-item interaction and rapidly drifting contextual features. Specifically, our proposed HySAGE network learns a relatively static graph embedding from user-item interaction and an adaptive embedding from drifting contextual features. These embeddings are incorporated into an interest network to generate the user interest in some certain context. We adopt an interactive attention module to learn the interactions among static graph embeddings, adaptive contextual embeddings, and user interest, helping to achieve a better final representation. Extensive experiments on real-world datasets demonstrate that HySAGE significantly improves the performance of the existing state-of-the-art recommendation algorithms.

cs.IR↗

arXiv2022

Eight Years of Face Recognition Research: Reproducibility, Achievements and Open Issues

Automatic face recognition is a research area with high popularity. Many different face recognition algorithms have been proposed in the last thirty years of intensive research in the field. With the popularity of deep learning and its capability to solve a huge variety of different problems, face recognition researchers have concentrated effort on creating better models under this paradigm. From the year 2015, state-of-the-art face recognition has been rooted in deep learning models. Despite the availability of large-scale and diverse datasets for evaluating the performance of face recognition algorithms, many of the modern datasets just combine different factors that influence face recognition, such as face pose, occlusion, illumination, facial expression and image quality. When algorithms produce errors on these datasets, it is not clear which of the factors has caused this error and, hence, there is no guidance in which direction more research is required. This work is a followup from our previous works developed in 2014 and eventually published in 2016, showing the impact of various facial aspects on face recognition algorithms. By comparing the current state-of-the-art with the best systems from the past, we demonstrate that faces under strong occlusions, some types of illumination, and strong expressions are problems mastered by deep learning algorithms, whereas recognition with low-resolution images, extreme pose variations, and open-set recognition is still an open problem. To show this, we run a sequence of experiments using six different datasets and five different face recognition algorithms in an open-source and reproducible manner. We provide the source code to run all of our experiments, which is easily extensible so that utilizing your own deep network in our evaluation is just a few minutes away.

cs.CV↗

arXiv2022

Axionic Dirac seesaw and electroweak vacuum stability

We explore the connection between tree-level Dirac neutrino masses and axion physics in a scenario where the PQ symmetry enforces lepton number conservation perturbatively. Requiring that the PQ scale $f_a$ is the only heavy scale to play a role in neutrino mass generation, we are led to the construction of a KSVZ-type model where Dirac neutrino masses are inversely proportional to $f_a$, provided a real scalar triplet (zero hypercharge) is added to the SM scalar sector. We analyse this extended scalar sector, focusing on the stabilisation of the electroweak vacuum. The contribution of the triplet VEV to the $W$ mass may also be responsible for the recent hint of beyond-the-SM physics by the CDF collaboration.

hep-ph↗

arXiv2022

FPIRM: Floating-point Processing in Racetrack Memories

Convolutional neural networks (CNN) have become a ubiquitous algorithm with growing applications in mobile and edge settings. We describe a compute-in-memory (CIM) technique called FPIRM using Racetrack Memory (RM) to accelerate CNNs for edge systems. Using transverse read, a technique that can determine the number of '1's multiple adjacent domains, FPIRM can efficiently implement multi-operand bulk-bitwise and addition computations, and two-operand multiplication. We discuss how FPIRM can implement both variable precision integer and floating point arithmetic. This allows both CNN inference and on-device training without expensive data movement to the cloud. Based on these functions we demonstrate implementation of several CNNs with back propagation using RM CIM and compare these to state-of-the-art implementations of CIM inference and training in Field-Programmable Gate Arrays. During training FPIRM improves by 2$\times$ the efficiency, by reducing the energy consumption by at least 27% and increasing the throughput by at least 18% against FPGA.

cs.ET↗

arXiv2022

H2H: Heterogeneous Model to Heterogeneous System Mapping with Computation and Communication Awareness

The complex nature of real-world problems calls for heterogeneity in both machine learning (ML) models and hardware systems. The heterogeneity in ML models comes from multi-sensor perceiving and multi-task learning, i.e., multi-modality multi-task (MMMT), resulting in diverse deep neural network (DNN) layers and computation patterns. The heterogeneity in systems comes from diverse processing components, as it becomes the prevailing method to integrate multiple dedicated accelerators into one system. Therefore, a new problem emerges: heterogeneous model to heterogeneous system mapping (H2H). While previous mapping algorithms mostly focus on efficient computations, in this work, we argue that it is indispensable to consider computation and communication simultaneously for better system efficiency. We propose a novel H2H mapping algorithm with both computation and communication awareness; by slightly trading computation for communication, the system overall latency and energy consumption can be largely reduced. The superior performance of our work is evaluated based on MAESTRO modeling, demonstrating 15%-74% latency reduction and 23%-64% energy reduction compared with existing computation-prioritized mapping algorithms.

cs.LG↗

arXiv2022

Towards Dynamic and Safe Configuration Tuning for Cloud Databases

Configuration knobs of database systems are essential to achieve high throughput and low latency. Recently, automatic tuning systems using machine learning methods (ML) have shown to find better configurations compared to experienced database administrators (DBAs). However, there are still gaps to apply the existing systems in production environments, especially in the cloud. First, they conduct tuning for a given workload within a limited time window and ignore the dynamicity of workloads and data. Second, they rely on a copied instance and do not consider the availability of the database when sampling configurations, making the tuning expensive, delayed, and unsafe. To fill these gaps, we propose OnlineTune, which tunes the online databases safely in changing cloud environments. To accommodate the dynamicity, OnlineTune embeds the environmental factors as context feature and adopts contextual Bayesian Optimization with context space partition to optimize the database adaptively and scalably. To pursue safety during tuning, we leverage the black-box and the white-box knowledge to evaluate the safety of configurations and propose a safe exploration strategy via subspace adaptation.%, greatly decreasing the risks of applying bad configurations. We conduct evaluations on dynamic workloads from benchmarks and real-world workloads. Compared with the state-of-the-art methods, OnlineTune achieves 14.4%~165.3% improvement on cumulative performance while reducing 91.0%~99.5% unsafe configuration recommendations.

cs.DB↗

arXiv2022

Facilitating Database Tuning with Hyper-Parameter Optimization: A Comprehensive Experimental Evaluation

Recently, using automatic configuration tuning to improve the performance of modern database management systems (DBMSs) has attracted increasing interest from the database community. This is embodied with a number of systems featuring advanced tuning capabilities being developed. However, it remains a challenge to select the best solution for database configuration tuning, considering the large body of algorithm choices. In addition, beyond the applications on database systems, we could find more potential algorithms designed for configuration tuning. To this end, this paper provides a comprehensive evaluation of configuration tuning techniques from a broader perspective, hoping to better benefit the database community. In particular, we summarize three key modules of database configuration tuning systems and conduct extensive ablation studies using various challenging cases. Our evaluation demonstrates that the hyper-parameter optimization algorithms can be borrowed to further enhance the database configuration tuning. Moreover, we identify the best algorithm choices for different modules. Beyond the comprehensive evaluations, we offer an efficient and unified database configuration tuning benchmark via surrogates that reduces the evaluation cost to a minimum, allowing for extensive runs and analysis of new techniques.

cs.DB↗

arXiv2022

QSDsan: An Integrated Platform for Quantitative Sustainable Design of Sanitation and Resource Recovery Systems

Sustainable sanitation and resource recovery technologies are needed to address rapid environmental and socioeconomic changes. Research prioritization is critical to expedite the development and deployment of such technologies across their vast system space (e.g., technology choices, design and operating decisions). In this study, we introduce QSDsan - an open-source tool written in Python (under the object-oriented programming paradigm) and developed for the quantitative sustainable design (QSD) of sanitation and resource recovery systems. As an integrated platform for system design, process modeling and simulation, techno-economic analysis (TEA), and life cycle assessment (LCA), QSDsan can be used to enumerate and investigate the opportunity space for emerging technologies under uncertainty, while considering contextual parameters that are critical to technology deployment. We illustrate the core capabilities of QSDsan through two distinct examples: (i) evaluation of a complete sanitation value chain that compares three alternative systems; and (ii) dynamic simulation of the wastewater treatment plant described in the benchmark simulation model no. 1 (BSM1). Through these examples, we show the utility of QSDsan to automate design, enable flexible process modeling, achieve rapid and reproducible simulations, and to perform advanced statistical analyses with integrated visualization. We strive to make QSDsan a community-led platform with online documentation, tutorials (explanatory notes, executable scripts, and video demonstrations), and a growing ecosystem of supporting packages (e.g., DMsan for decision-making). This platform can be freely accessed, used, and expanded by researchers, practitioners, and the public alike, ultimately contributing to the advancement of safe and affordable sanitation technologies around the globe.

cs.CY↗

arXiv2022

A Topological Solution of Entanglement for Complex-shaped Parts in Robotic Bin-picking

This paper addresses the problem of picking up only one object at a time avoiding any entanglement in bin-picking. To cope with a difficult case where the complex-shaped objects are heavily entangled together, we propose a topology-based method that can generate non-tangle grasp positions on a single depth image. The core technique is entanglement map, which is a feature map to measure the entanglement possibilities obtained from the input image. We use the entanglement map to select probable regions containing graspable objects. The optimum grasping pose is detected from the selected regions considering the collision between robot hand and objects. Experimental results show that our analytic method provides a more comprehensive and intuitive observation of entanglement and exceeds previous learning-based work in success rates. Especially, our topology-based method does not rely on any object models or time-consuming training process, so that it can be easily adapted to more complex bin-picking scenes.

cs.RO↗

arXiv2022

EF-Train: Enable Efficient On-device CNN Training on FPGA Through Data Reshaping for Online Adaptation or Personalization

Conventionally, DNN models are trained once in the cloud and deployed in edge devices such as cars, robots, or unmanned aerial vehicles (UAVs) for real-time inference. However, there are many cases that require the models to adapt to new environments, domains, or new users. In order to realize such domain adaption or personalization, the models on devices need to be continuously trained on the device. In this work, we design EF-Train, an efficient DNN training accelerator with a unified channel-level parallelism-based convolution kernel that can achieve end-to-end training on resource-limited low-power edge-level FPGAs. It is challenging to implement on-device training on resource-limited FPGAs due to the low efficiency caused by different memory access patterns among forward, backward propagation, and weight update. Therefore, we developed a data reshaping approach with intra-tile continuous memory allocation and weight reuse. An analytical model is established to automatically schedule computation and memory resources to achieve high energy efficiency on edge FPGAs. The experimental results show that our design achieves 46.99 GFLOPS and 6.09GFLOPS/W in terms of throughput and energy efficiency, respectively.

cs.LG↗

arXiv2022

Defense Strategies Toward Model Poisoning Attacks in Federated Learning: A Survey

Advances in distributed machine learning can empower future communications and networking. The emergence of federated learning (FL) has provided an efficient framework for distributed machine learning, which, however, still faces many security challenges. Among them, model poisoning attacks have a significant impact on the security and performance of FL. Given that there have been many studies focusing on defending against model poisoning attacks, it is necessary to survey the existing work and provide insights to inspire future research. In this paper, we first classify defense mechanisms for model poisoning attacks into two categories: evaluation methods for local model updates and aggregation methods for the global model. Then, we analyze some of the existing defense strategies in detail. We also discuss some potential challenges and future research directions. To the best of our knowledge, we are the first to survey defense methods for model poisoning attacks in FL.

cs.CR↗

arXiv2022

Modified gravity models for inflation: In conformity with observations

We consider a modified gravity framework for inflation by adding to the Einstein-Hilbert action a direct $f(ϕ)T$ term, where $ϕ$ is identified as the inflaton and $T$ is the trace of the energy-momentum tensor. The framework goes to Einstein gravity naturally when the inflaton decays out. We investigate inflation dynamics in this $f(ϕ)T$ gravity (not to be confused with torsion-scalar coupled theories) on a general basis and then apply it to three well-motivated inflationary models. We find that the predictions for the spectral tilt and the tensor-to-scalar ratio are sensitive to this new $f(ϕ)T$ term. This $f(ϕ)T$ gravity brings chaotic and natural inflation into better agreement with data and allows a larger tensor-to-scalar ratio in the Starobinsky model.

gr-qc↗

arXiv2021

Speech Technology for Everyone: Automatic Speech Recognition for Non-Native English with Transfer Learning

To address the performance gap of English ASR models on L2 English speakers, we evaluate fine-tuning of pretrained wav2vec 2.0 models (Baevski et al., 2020; Xu et al., 2021) on L2-ARCTIC, a non-native English speech corpus (Zhao et al., 2018) under different training settings. We compare \textbf{(a)} models trained with a combination of diverse accents to ones trained with only specific accents and \textbf{(b)} results from different single-accent models. Our experiments demonstrate the promise of developing ASR models for non-native English speakers, even with small amounts of L2 training data and even without a language model. Our models also excel in the zero-shot setting where we train on multiple L2 datasets and test on a blind L2 test set.

eess.AS↗

arXiv2021

Inverse Seesaw Model with a Modular $S^{}_4$ Symmetry: Lepton Flavor Mixing and Warm Dark Matter

In this paper, we present a systematic investigation on simple inverse seesaw models for neutrino masses and flavor mixing based on the modular $S^{}_4$ symmetry. Two right-handed neutrinos and three extra fermion singlets are introduced to account for light neutrino masses through the inverse seesaw mechanism, and to provide a keV-mass sterile neutrino as the candidate for warm dark matter in our Universe. Considering all possible modular forms with weights no larger than four, we obtain twelve models, among which we find one is in excellent agreement with the observed lepton mass spectra and flavor mixing. Moreover, we explore the allowed range of the sterile neutrino mass and mixing angles, by taking into account the direct search of $X$-ray line and the Lyman-$α$ observations. The model predictions for neutrino mixing parameters and the dark matter abundance will be readily testable in future neutrino oscillation experiments and cosmological observations.

hep-ph↗

arXiv2021

Warm-assisted natural inflation

We consider natural inflation in a warm inflation framework with a temperature-dependent dissipative coefficient $Γ\propto T^3$. Natural inflation can be compatible with the Planck 2018 results with such warm assistance. With no a priori assumptions on the dissipative effect's magnitude, we find that the Planck results prefer a weak dissipative regime for our benchmark scale $f=5 M_{\rm pl}$, which lies outside the $2σ$ region in the cold case. The inflation starts in the cold regime and evolves with a growing thermal fluctuation that dominates over quantum fluctuation before the end of the inflation. The observed spectral tilt puts stringent constraints on the model's parameter space. We find that $f< 1 M_{\rm pl}$ is excluded. A possible origin of such dissipative coefficient from axion-like coupling to gauge fields and tests of the model are also discussed.

astro-ph.CO↗

arXiv2021

Thief, Beware of What Get You There: Towards Understanding Model Extraction Attack

Model extraction increasingly attracts research attentions as keeping commercial AI models private can retain a competitive advantage. In some scenarios, AI models are trained proprietarily, where neither pre-trained models nor sufficient in-distribution data is publicly available. Model extraction attacks against these models are typically more devastating. Therefore, in this paper, we empirically investigate the behaviors of model extraction under such scenarios. We find the effectiveness of existing techniques significantly affected by the absence of pre-trained models. In addition, the impacts of the attacker's hyperparameters, e.g. model architecture and optimizer, as well as the utilities of information retrieved from queries, are counterintuitive. We provide some insights on explaining the possible causes of these phenomena. With these observations, we formulate model extraction attacks into an adaptive framework that captures these factors with deep reinforcement learning. Experiments show that the proposed framework can be used to improve existing techniques, and show that model extraction is still possible in such strict scenarios. Our research can help system designers to construct better defense strategies based on their scenarios.

cs.LG↗