Weiming Li

Jinlai Zhang, Weiming Li, Shuang Liang et al.

Estimating 6D object pose from an RGB image is important for many real-world applications such as autonomous driving and robotic grasping. Recent deep learning models have achieved significant progress on this task but their robustness received little research attention. In this work, for the first time, we study adversarial samples that can fool deep learning models with imperceptible perturbations to input image. In particular, we propose a Unified 6D pose estimation Attack, namely U6DA, which can successfully attack several state-of-the-art (SOTA) deep learning models for 6D pose estimation. The key idea of our U6DA is to fool the models to predict wrong results for object instance localization and shape that are essential for correct 6D pose estimation. Specifically, we explore a transfer-based black-box attack to 6D pose estimation. We design the U6DA loss to guide the generation of adversarial examples, the loss aims to shift the segmentation attention map away from its original position. We show that the generated adversarial samples are not only effective for direct 6D pose estimation models, but also are able to attack two-stage models regardless of their robust RANSAC modules. Extensive experiments were conducted to demonstrate the effectiveness, transferability, and anti-defense capability of our U6DA on large-scale public benchmarks. We also introduce a new U6DA-Linemod dataset for robustness study of the 6D pose estimation task. Our codes and dataset will be available at \url{https://github.com/cuge1995/U6DA}.

Weiming Li, Manni Duan, Dong An et al.

Large language models (LLMs) demonstrate extraordinary abilities in a wide range of natural language processing (NLP) tasks. In this paper, we show that, beyond text understanding capability, LLMs are capable of processing text layouts that are denoted by spatial markers. They are able to answer questions that require explicit spatial perceiving and reasoning, while a drastic performance drop is observed when the spatial markers from the original data are excluded. We perform a series of experiments with the GPT-3.5, Baichuan2, Llama2 and ChatGLM3 models on various types of layout-sensitive datasets for further analysis. The experimental results reveal that the layout understanding ability of LLMs is mainly introduced by the coding data for pretraining, which is further enhanced at the instruction-tuning stage. In addition, layout understanding can be enhanced by integrating low-cost, auto-generated data approached by a novel text game. Finally, we show that layout understanding ability is beneficial for building efficient visual question-answering (VQA) systems.

Wei-Ming Li, Shi-Ju Ran

In quantum and quantum-inspired machine learning, the very first step is to embed the data in quantum space known as Hilbert space. Developing quantum kernel function (QKF), which defines the distances among the samples in the Hilbert space, belongs to the fundamental topics for machine learning. In this work, we propose the rescaled logarithmic fidelity (RLF) and non-parametric semi-supervised learning in the quantum space, which we name as RLF-NSSL. The rescaling takes advantage of the non-linearity of the kernel to tune the mutual distances of samples in the Hilbert space, and meanwhile avoids the exponentially-small fidelities between quantum many-qubit states. Being non-parametric excludes the possible effects from the variational parameters, and evidently demonstrates the advantages from the space itself. We compare RLF-NSSL with several well-known non-parametric algorithms including naive Bayes classifiers, k-nearest neighbors, and spectral clustering. Our method exhibits better accuracy particularly for the unsupervised case with no labeled samples and the few-shot cases with small numbers of labeled samples. With the visualizations by t-stochastic neighbor embedding, our results imply that the machine learning in the Hilbert space complies with the principles of maximal coding rate reduction, where the low-dimensional data exhibit within-class compressibility, between-class discrimination, and overall diversity. Our proposals can be applied to other quantum and quantum-inspired machine learning, including the methods using the parametric models such as tensor networks, quantum circuits, and quantum neural networks.