Yi Li

Hualiang Wang, Yi Li, Huifeng Yao et al.

Out-of-distribution (OOD) detection refers to training the model on an in-distribution (ID) dataset to classify whether the input images come from unknown classes. Considerable effort has been invested in designing various OOD detection methods based on either convolutional neural networks or transformers. However, zero-shot OOD detection methods driven by CLIP, which only require class names for ID, have received less attention. This paper presents a novel method, namely CLIP saying no (CLIPN), which empowers the logic of saying no within CLIP. Our key motivation is to equip CLIP with the capability of distinguishing OOD and ID samples using positive-semantic prompts and negation-semantic prompts. Specifically, we design a novel learnable no prompt and a no text encoder to capture negation semantics within images. Subsequently, we introduce two loss functions: the image-text binary-opposite loss and the text semantic-opposite loss, which we use to teach CLIPN to associate images with no prompts, thereby enabling it to identify unknown samples. Furthermore, we propose two threshold-free inference algorithms to perform OOD detection by utilizing negation semantics from no prompts and the text encoder. Experimental results on 9 benchmark datasets (3 ID datasets and 6 OOD datasets) for the OOD detection task demonstrate that CLIPN, based on ViT-B-16, outperforms 7 well-used algorithms by at least 2.34% and 11.64% in terms of AUROC and FPR95 for zero-shot OOD detection on ImageNet-1K. Our CLIPN can serve as a solid foundation for effectively leveraging CLIP in downstream OOD tasks. The code is available on https://github.com/xmed-lab/CLIPN.

Zhenyi Ji, Wenyuan Wu, Yi Li et al.

In this paper, based on the homotopy continuation method and the interval Newton method, an efficient algorithm is introduced to isolate the real roots of semi-algebraic system. Tests on some random examples and a variety of problems including transcendental functions arising in many applications show that the new algorithm reduces the cost substantially compared with the traditional symbolic approaches.

Victor DeCaria, Ahmet Guzel, William Layton et al.

Variable Stepsize Variable Order (VSVO) methods are the methods of choice to efficiently solve a wide range of ODEs with minimal work and assured accuracy. However, VSVO methods have limited impact in timestepping methods in complex applications due to their computational complexity and the difficulty to implement them in legacy code. We introduce a family of implicit, embedded, VSVO methods that require only one BDF solve at each time step followed by adding linear combinations of the solution at previous time levels. In particular, we construct implicit and linearly implicit VSVO methods of orders two, three and four with the same computational complexity as variable stepsize BDF3. The choice of changing the order of the method is simple and does not require additional solves of linear or nonlinear systems.