It is a challenging task to learn rich and multi-scale spatiotemporal semantics from high-dimensional videos, due to large local redundancy and complex global dependency between video frames. The recent advances in this research have been mainly driven by 3D convolutional neural networks and vision transformers. Although 3D convolution can efficiently aggregate local context to suppress local redundancy from a small 3D neighborhood, it lacks the capability to capture global dependency because of the limited receptive field. Alternatively, vision transformers can effectively capture long-range dependency by self-attention mechanism, while having the limitation on reducing local redundancy with blind similarity comparison among all the tokens in each layer. Based on these observations, we propose a novel Unified transFormer (UniFormer) which seamlessly integrates merits of 3D convolution and spatiotemporal self-attention in a concise transformer format, and achieves a preferable balance between computation and accuracy. Different from traditional transformers, our relation aggregator can tackle both spatiotemporal redundancy and dependency, by learning local and global token affinity respectively in shallow and deep layers. We conduct extensive experiments on the popular video benchmarks, e.g., Kinetics-400, Kinetics-600, and Something-Something V1&V2. With only ImageNet-1K pretraining, our UniFormer achieves 82.9%/84.8% top-1 accuracy on Kinetics-400/Kinetics-600, while requiring 10 fewer GFLOPs than other state-of-the-art methods. For Something-Something V1 and V2, our UniFormer achieves new state-of-the-art performances of 60.9% and 71.2% top-1 accuracy respectively. Code is available at https://github.com/Sense-X/UniFormer.

Accurate protein side-chain modeling is crucial for protein folding and protein design. In the past decades, many successful methods have been proposed to address this issue. However, most of them depend on the discrete samples from the rotamer library, which may have limitations on their accuracies and usages. In this study, we report an open-source toolkit for protein side-chain modeling, named OPUS-Rota4. It consists of three modules: OPUS-RotaNN2, which predicts protein side-chain dihedral angles; OPUS-RotaCM, which measures the distance and orientation information between the side chain of different residue pairs and OPUS-Fold2, which applies the constraints derived from the first two modules to guide side-chain modeling. OPUS-Rota4 adopts the dihedral angles predicted by OPUS-RotaNN2 as its initial states, and uses OPUS-Fold2 to refine the side-chain conformation with the side-chain contact map constraints derived from OPUS-RotaCM. Therefore, we convert the side-chain modeling problem into a side-chain contact map prediction problem. OPUS-Fold2 is written in Python and TensorFlow2.4, which is user-friendly to include other differentiable energy terms. OPUS-Rota4 also provides a platform in which the side-chain conformation can be dynamically adjusted under the influence of other processes. We apply OPUS-Rota4 on 15 FM predictions submitted by AlphaFold2 on CASP14, the results show that the side chains modeled by OPUS-Rota4 are closer to their native counterparts than those predicted by AlphaFold2 (e.g. the residue-wise RMSD for all residues and core residues are 0.588 and 0.472 for AlphaFold2, and 0.535 and 0.407 for OPUS-Rota4).

Enormous waves of technological innovations over the past several years, marked by the advances in AI technologies, are profoundly reshaping the industry and the society. However, down the road, a key challenge awaits us, that is, our capability of meeting rapidly-growing scenario-specific demands is severely limited by the cost of acquiring a commensurate amount of training data. This difficult situation is in essence due to limitations of the mainstream learning paradigm: we need to train a new model for each new scenario, based on a large quantity of well-annotated data and commonly from scratch. In tackling this fundamental problem, we move beyond and develop a new learning paradigm named INTERN. By learning with supervisory signals from multiple sources in multiple stages, the model being trained will develop strong generalizability. We evaluate our model on 26 well-known datasets that cover four categories of tasks in computer vision. In most cases, our models, adapted with only 10% of the training data in the target domain, outperform the counterparts trained with the full set of data, often by a significant margin. This is an important step towards a promising prospect where such a model with general vision capability can dramatically reduce our reliance on data, thus expediting the adoption of AI technologies. Furthermore, revolving around our new paradigm, we also introduce a new data system, a new architecture, and a new benchmark, which, together, form a general vision ecosystem to support its future development in an open and inclusive manner.

​Large-scale contrastive vision-language pretraining has shown significant progress in visual representation learning. Unlike traditional visual systems trained by a fixed set of discrete labels, a new paradigm was introduced in (Radford et al., 2021) to directly learn to align images with raw texts in an open-vocabulary setting. On downstream tasks, a carefully chosen text prompt is employed to make zero-shot predictions. To avoid non-trivial prompt engineering, context optimization (Zhou et al., 2021) has been proposed to learn continuous vectors as task-specific prompts with few-shot training examples. In this paper, we show that there is an alternative path to achieve better vision-language models other than prompt tuning. While prompt tuning is for the textual inputs, we propose CLIP-Adapter to conduct fine-tuning with feature adapters on either visual or language branch. Specifically, CLIP-Adapter adopts an additional bottleneck layer to learn new features and performs residual style feature blending with the original pre-trained features. As a consequence, CLIP-Adapter is able to outperform context optimization while maintains a simple design. Experiments and extensive ablation studies on various visual classification tasks demonstrate the effectiveness of our approach.

Convolutional neural networks (CNNs) are ubiquitous in computer vision, with a myriad of effective and efficient variations. Recently, Transformers – originally introduced in natural language processing – have been increasingly adopted in computer vision. While early adopters continue to employ CNN backbones, the latest networks are end-to-end CNN-free Transformer solutions. A recent surprising finding shows that a simple MLP based solution without any traditional convolutional or Transformer components can produce effective visual representations. While CNNs, Transformers and MLP-Mixers may be considered as completely disparate architectures, we provide a unified view showing that they are in fact special cases of a more general method to aggregate spatial context in a neural network stack. We present the CONTAINER (CONText AggregatIon NEtwoRk), a general-purpose building block for multi-head context aggregation that can exploit long-range interactions a la Transformers while still exploiting the inductive bias of the local convolution operation leading to faster convergence speeds, often seen in CNNs. Our CONTAINER architecture achieves 82.7 % Top-1 accuracy on ImageNet using 22M parameters, +2.8 improvement compared with DeiT-Small, and can converge to 79.9 % Top-1 accuracy in just 200 epochs. In contrast to Transformer-based methods that do not scale well to downstream tasks that rely on larger input image resolutions, our efficient network, named CONTAINER-LIGHT, can be employed in object detection and instance segmentation networks such as DETR, RetinaNet and Mask-RCNN to obtain an impressive detection mAP of 38.9, 43.8, 45.1 and mask mAP of 41.3, providing large improvements of 6.6, 7.3, 6.9 and 6.6 pts respectively, compared to a ResNet-50 backbone with a comparable compute and parameter size. Our method also achieves promising results on selfsupervised learning compared to DeiT on the DINO framework. Code is released at https://github.com/allenai/container.

The recently proposed Detection Transformer (DETR)model successfully applies Transformer to objects detection and achieves comparable performance with two-stage object detection frameworks, such as Faster-RCNN. However, DETR suffers from its slow convergence. Training DETR [4] from scratch needs 500 epochs to achieve a high accuracy. To accelerate its convergence, we propose a simple yet effective scheme for improving the DETR framework, namely Spatially Modulated Co-Attention (SMCA) mechanism. The core idea of SMCA is to conduct locationaware co-attention in DETR by constraining co-attention responses to be high near initially estimated bounding box locations. Our proposed SMCA increases DETR’s convergence speed by replacing the original co-attention mechanism in the decoder while keeping other operations in DETR unchanged. Furthermore, by integrating multi-head and scale-selection attention designs into SMCA, our fully fledged SMCA can achieve better performance compared to DETR with a dilated convolution-based backbone (45.6 mAP at 108 epochs vs. 43.3 mAP at 500 epochs). We perform extensive ablation studies on COCO dataset to validate SMCA.