no code implementations • NeurIPS 2023 • Weijie Tu, Weijian Deng, Tom Gedeon
Driven by the above, this work comprehensively investigates the safety objectives of CLIP models, specifically focusing on three key properties: resilience to visual factor variations, calibrated uncertainty estimations, and the ability to detect anomalous inputs.
no code implementations • 12 Feb 2024 • Weijie Tu, Weijian Deng, Dylan Campbell, Stephen Gould, Tom Gedeon
Vision--Language Models (VLMs) have emerged as the dominant approach for zero-shot recognition, adept at handling diverse scenarios and significant distribution changes.
no code implementations • 19 Oct 2023 • Chunyi Sun, Junlin Han, Weijian Deng, Xinlong Wang, Zishan Qin, Stephen Gould
Our work highlights the potential of LLMs in 3D modeling, offering a basic framework for future advancements in scene generation and animation.
no code implementations • CVPR 2023 • Weijie Tu, Weijian Deng, Tom Gedeon, Liang Zheng
The former measures how suitable a training set is for a target domain, while the latter studies how challenging a test set is for a learned model.
no code implementations • 9 Mar 2023 • Yuli Zou, Weijian Deng, Liang Zheng
In other words, a calibrator optimal on the calibration set would be suboptimal on the OOD test set and thus has degraded performance.
no code implementations • 2 Feb 2023 • Weijian Deng, Yumin Suh, Stephen Gould, Liang Zheng
This work aims to assess how well a model performs under distribution shifts without using labels.
1 code implementation • ICCV 2023 • Yuli Zou, Weijian Deng, Liang Zheng
With this knowledge, we propose a simple and effective method named adaptive calibrator ensemble (ACE) to calibrate OOD datasets whose difficulty is usually higher than the calibration set.
no code implementations • 14 Jul 2022 • Weijian Deng, Stephen Gould, Liang Zheng
Generalization and invariance are two essential properties of any machine learning model.
2 code implementations • ICCV 2021 • Xiaoxiao Sun, Yunzhong Hou, Weijian Deng, Hongdong Li, Liang Zheng
For this problem, we propose to adopt a proxy dataset that 1) is fully labeled and 2) well reflects the true model rankings in a given target environment, and use the performance rankings on the proxy sets as surrogates.
no code implementations • 10 Jun 2021 • Weijian Deng, Stephen Gould, Liang Zheng
In this work, we train semantic classification and rotation prediction in a multi-task way.
no code implementations • 12 Dec 2020 • Weijian Deng, Joshua Marsh, Stephen Gould, Liang Zheng
The memory module stores the prototypical feature representation for each category as a moving average.
Ranked #50 on Fine-Grained Image Classification on CUB-200-2011
no code implementations • CVPR 2021 • Weijian Deng, Liang Zheng
As the classification accuracy of the model on each sample (dataset) is known from the original dataset labels, our task can be solved via regression.
no code implementations • 3 Dec 2018 • Weijian Deng, Liang Zheng, Jianbin Jiao
When aligning the distributions in the embedding space, SCA enforces a similarity-preserving constraint to maintain class-level relations among the source and target images, i. e., if a source image and a target image are of the same class label, their corresponding embeddings are supposed to be aligned nearby, and vise versa.
no code implementations • 26 Nov 2018 • Weijian Deng, Liang Zheng, Qixiang Ye, Yi Yang, Jianbin Jiao
It first preserves two types of unsupervised similarity, namely, self-similarity of an image before and after translation, and domain-dissimilarity of a translated source image and a target image.
2 code implementations • CVPR 2018 • Weijian Deng, Liang Zheng, Qixiang Ye, Guoliang Kang, Yi Yang, Jianbin Jiao
To this end, we propose to preserve two types of unsupervised similarities, 1) self-similarity of an image before and after translation, and 2) domain-dissimilarity of a translated source image and a target image.
no code implementations • ICCV 2017 • Yifan Sun, Liang Zheng, Weijian Deng, Shengjin Wang
This paper proposes the SVDNet for retrieval problems, with focus on the application of person re-identification (re-ID).
Ranked #14 on Person Re-Identification on CUHK03 detected