A crucial aspect of understanding the complex nature of Deep Neural Networks (DNNs) is the ability to explain learned concepts within their latent representations. While methods exist to connect neurons to human-understandable textual descriptions, evaluating the quality of these explanations is challenging due to the lack of a unified quantitative approach.

Score distillation sampling (SDS) has proven to be an important tool, enabling the use of large-scale diffusion priors for tasks operating in data-poor domains. Unfortunately, SDS has a number of characteristic artifacts that limit its usefulness in general-purpose applications. In this paper, we make progress toward understanding the behavior of SDS and its variants by viewing them as solving an optimal-cost transport path from a source distribution to a target distribution. Under this new interpretation, these methods seek to transport corrupted images (source) to the natural image distribution (target). We argue that current methods' characteristic artifacts are caused by (1) linear approximation of the optimal path and (2) poor estimates of the source distribution. We show that calibrating the text conditioning of the source distribution can produce high-quality generation and translation results with little extra overhead. Our method can be easily applied across many domains, matching or beating the performance of specialized methods. We demonstrate its utility in text-to-2D, text-based NeRF optimization, translating paintings to real images, optical illusion generation, and 3D sketch-to-real. We compare our method to existing approaches for score distillation sampling and show that it can produce high-frequency details with realistic colors.

Unpaired image-to-image translation aims to translate an input image to another domain such that the output image looks like an image from another domain while important semantic information are preserved. Inferring the optimal mapping with unpaired data is impossible without making any assumptions. In this paper, we make a density changing assumption where image patches of high probability density should be mapped to patches of high probability density in another domain. Then we propose an efficient way to enforce this assumption: we train the flows as density estimators and penalize the variance of density changes. Despite its simplicity, our method achieves the best performance on benchmark datasets and needs only 56 86% of training time of the existing state-of-the-art method. The training and evaluation code are avaliable at https://github.com/Mid-Push/

Recent action recognition models have achieved impressive results by integrating objects, their locations and interactions. However, obtaining dense structured annotations for each frame is tedious and time-consuming, making these methods expensive to train and less scalable. On the other hand, one does often have access to a small set of annotated images, either within or outside the domain of interest. Here we ask how such images can be leveraged for downstream video understanding tasks. We propose a learning framework StructureViT (SViT for short), which demonstrates how utilizing the structure of a small number of images only available during training can improve a video model.

Table 6 provides summary statistics of domain coverage. Overall, the benchmark covers 8,637 biology images and 8,678 pathology images across 12 subdomains. Similarly, Table 7 shows summary statistics of microscopy modalities covered by Micro-Bench perception, including 10,864 images for light microscopy, 5,618 for fluorescence microscopy, and 833 images for electron microscopy across 8 microscopy imaging submodalities and 25 unique microscopy staining techniques (see Table 8). Micro-Bench Perception (Coarse-grained): Hierarchical metadata for each of the 17,235 perception images and task-specific templates (shown in Table 23) are used to create 5 coarse-grained questions and captions regarding microscopy modality, submodality, domain, subdomain, and staining technique. The use of hierarchical metadata enables the generation of options within each hierarchical level.

These settings are consistent with CAT-R and CAT-A. For CAT-R-2, we apply regular-Rwin, and set [sw, sh] as [4, 16] (same as CAT-R). We set the MLP expansion ratio as 2, consistent with SwinIR [13]. For CAT-A-2, we apply axial-Rwin, and set sl as 4 for all CATB in each RG. The MLP expansion ratio is set as 4. Best and second best results are colored with red and blue.

The best results are highlighted by bold. It can be clearly seen that our alternating reverse filtering network performs the best compared with other state-of-the-art methods in all the indexes, indicating the superiority of our proposed method. Images in the last row are the MSE residues between the fused results and the ground truth. Compared with other competing methods, our model has minor spatial and spectral distortions. It can be easily concluded from the observation of MSE maps.

Model-based methods produce coarse Co-SOD results due to hand-crafted intra-and inter-saliency features. Current data-driven models exploit inter-saliency cues, but undervalue the potential power of intra-saliency cues. In this paper, we propose an Intra-saliency Correlation Network (ICNet) to extract intra-saliency cues from the single image saliency maps (SISMs) predicted by any off-the-shelf SOD method, and obtain inter-saliency cues by correlation techniques. Specifically, we adopt normalized masked average pooling (NMAP) to extract latent intra-saliency categories from the SISMs and semantic features as intra cues. Then we employ a correlation fusion module (CFM) to obtain inter cues by exploiting correlations between the intra cues and single-image features. To improve Co-SOD performance, we propose a category-independent rearranged self-correlation feature (RSCF) strategy. Experiments on three benchmarks show that our ICNet outperforms previous state-of-the-art methods on Co-SOD.

Recently, AutoRegressive (AR) models for the whole image generation empowered by transformers have achieved comparable or even better performance compared to Generative Adversarial Networks (GANs). Unfortunately, directly applying such AR models to edit/change local image regions, may suffer from the problems of missing global information, slow inference speed, and information leakage of local guidance. To address these limitations, we propose a novel model - image Local Autoregressive Transformer (iLAT), to better facilitate the locally guided image synthesis. Our iLAT learns the novel local discrete representations, by the newly proposed local autoregressive (LA) transformer of the attention mask and convolution mechanism. Thus iLAT can efficiently synthesize the local image regions by key guidance information. Our iLAT is evaluated on various locally guided image syntheses, such as pose-guided person image synthesis and face editing. Both quantitative and qualitative results show the efficacy of our model.

We investigate the robustness of vision transformers (ViTs) through the lens of their special patch-based architectural structure, i.e., they process an image as a sequence of image patches. We find that ViTs are surprisingly insensitive to patchbased transformations, even when the transformation largely destroys the original semantics and makes the image unrecognizable by humans. This indicates that ViTs heavily use features that survived such transformations but are generally not indicative of the semantic class to humans. Further investigations show that these features are useful but non-robust, as ViTs trained on them can achieve high in-distribution accuracy, but break down under distribution shifts. From this understanding, we ask: can training the model to rely less on these features improve ViT robustness and out-of-distribution performance? We use the images transformed with our patch-based operations as negatively augmented views and offer losses to regularize the training away from using non-robust features. This is a complementary view to existing research that mostly focuses on augmenting inputs with semantic-preserving transformations to enforce models' invariance. We show that patch-based negative augmentation consistently improves robustness of ViTs on ImageNet based robustness benchmarks across 20+ different experimental settings. Furthermore, we find our patch-based negative augmentation are complementary to traditional (positive) data augmentation techniques and batchbased negative examples in contrastive learning.