Vision transformers have recently received explosive popularity, but the huge computational cost is still a severe issue. Recent efficient designs for vision transformers follow two pipelines, namely, structural compression based on local spatial prior and non-structural token pruning. However, token pruning breaks the spatial structure that is indispensable for local spatial prior. To take advantage of both two pipelines, this work seeks to dynamically identify uninformative tokens for each instance and trim down both the training and inference complexity while maintaining complete spatial structure and information flow. To achieve this goal, we propose Evo-ViT, a self-motivated slow-fast token evolution method for vision transformers. Specifically, we conduct unstructured instance-wise token selection by taking advantage of the global class attention that is unique to vision transformers. Then, we propose to update informative tokens and placeholder tokens that contribute little to the final prediction with different computational priorities, namely, slow-fast updating. Thanks to the slow-fast updating mechanism that guarantees information flow and spatial structure, our Evo-ViT can accelerate vanilla transformers of both flat and deep-narrow structures from the very beginning of the training process. Experimental results demonstrate that the proposed method can significantly reduce the computational costs of vision transformers while maintaining comparable performance on image classification. For example, our method accelerates DeiTS by over 60% throughput while only sacrificing 0.4% top-1 accuracy.
In this paper, we detail the relationship between convolutions and self-attention in natural language tasks. We show that relative position embeddings in self-attention layers are equivalent to recently-proposed dynamic lightweight convolutions, and we consider multiple new ways of integrating convolutions into Transformer self-attention. Specifically, we propose composite attention, which unites previous relative position embedding methods under a convolutional framework. We conduct experiments by training BERT with composite attention, finding that convolutions consistently improve performance on multiple downstream tasks, replacing absolute position embeddings. To inform future work, we present results comparing lightweight convolutions, dynamic convolutions, and depthwise-separable convolutions in language model pre-training, considering multiple injection points for convolutions in self-attention layers.
Big progress has been achieved in domain adaptation in decades. Existing works are always based on an ideal assumption that testing target domain are i.i.d. with training target domains. However, due to unpredictable corruptions (e.g., noise and blur) in real data like web images, domain adaptation methods are increasingly required to be corruption robust on target domains. In this paper, we investigate a new task, Corruption-agnostic Robust Domain Adaptation (CRDA): to be accurate on original data and robust against unavailable-for-training corruptions on target domains. This task is non-trivial due to large domain discrepancy and unsupervised target domains. We observe that simple combinations of popular methods of domain adaptation and corruption robustness have sub-optimal CRDA results. We propose a new approach based on two technical insights into CRDA: 1) an easy-to-plug module called Domain Discrepancy Generator (DDG) that generates samples that enlarge domain discrepancy to mimic unpredictable corruptions; 2) a simple but effective teacher-student scheme with contrastive loss to enhance the constraints on target domains. Experiments verify that DDG keeps or even improves performance on original data and achieves better corruption robustness that baselines.
In this paper, we present a regression-based pose recognition method using cascade Transformers. One way to categorize the existing approaches in this domain is to separate them into 1). heatmap-based and 2). regression-based. In general, heatmap-based methods achieve higher accuracy but are subject to various heuristic designs (not end-to-end mostly), whereas regression-based approaches attain relatively lower accuracy but they have less intermediate non-differentiable steps. Here we utilize the encoder-decoder structure in Transformers to perform regression-based person and keypoint detection that is general-purpose and requires less heuristic design compared with the existing approaches. We demonstrate the keypoint hypothesis (query) refinement process across different self-attention layers to reveal the recursive self-attention mechanism in Transformers. In the experiments, we report competitive results for pose recognition when compared with the competing regression-based methods.
In this paper, we present Co-scale conv-attentional image Transformers (CoaT), a Transformer-based image classifier equipped with co-scale and conv-attentional mechanisms. First, the co-scale mechanism maintains the integrity of Transformers' encoder branches at individual scales, while allowing representations learned at different scales to effectively communicate with each other; we design a series of serial and parallel blocks to realize the co-scale attention mechanism. Second, we devise a conv-attentional mechanism by realizing a relative position embedding formulation in the factorized attention module with an efficient convolution-like implementation. CoaT empowers image Transformers with enriched multi-scale and contextual modeling capabilities. On ImageNet, relatively small CoaT models attain superior classification results compared with the similar-sized convolutional neural networks and image/vision Transformers. The effectiveness of CoaT's backbone is also illustrated on object detection and instance segmentation, demonstrating its applicability to the downstream computer vision tasks.
Takagi-Sugeno-Kang (TSK) fuzzy system with Gaussian membership functions (MFs) is one of the most widely used fuzzy systems in machine learning. However, it usually has difficulty handling high-dimensional datasets. This paper explores why TSK fuzzy systems with Gaussian MFs may fail on high-dimensional inputs. After transforming defuzzification to an equivalent form of softmax function, we find that the poor performance is due to the saturation of softmax. We show that two defuzzification operations, LogTSK and HTSK, the latter of which is first proposed in this paper, can avoid the saturation. Experimental results on datasets with various dimensionalities validated our analysis and demonstrated the effectiveness of LogTSK and HTSK.
In this paper, we present a holistically end-to-end algorithm for line segment detection with transformers that is post-processing and heuristics-guided intermediate processing (edge/junction/region detection) free. Our method, named LinE segment TRansformers (LETR), tackles the three main problems in this domain, namely edge element detection, perceptual grouping, and holistic inference by three highlights in detection transformers (DETR) including tokenized queries with integrated encoding and decoding, self-attention, and joint queries respectively. The transformers learn to progressively refine line segments through layers of self-attention mechanism skipping the heuristic design in the previous line segmentation algorithms. We equip multi-scale encoder/decoder in the transformers to perform fine-grained line segment detection under a direct end-point distance loss that is particularly suitable for entities such as line segments that are not conveniently represented by bounding boxes. In the experiments, we show state-of-the-art results on Wireframe and YorkUrban benchmarks. LETR points to a promising direction for joint end-to-end detection of general entities beyond the standard object bounding box representation.
Several scientific studies have reported the existence of the income gap among rideshare drivers based on demographic factors such as gender, age, race, etc. In this paper, we study the income inequality among rideshare drivers due to discriminative cancellations from riders, and the tradeoff between the income inequality (called fairness objective) with the system efficiency (called profit objective). We proposed an online bipartite-matching model where riders are assumed to arrive sequentially following a distribution known in advance. The highlight of our model is the concept of acceptance rate between any pair of driver-rider types, where types are defined based on demographic factors. Specially, we assume each rider can accept or cancel the driver assigned to her, each occurs with a certain probability which reflects the acceptance degree from the rider type towards the driver type. We construct a bi-objective linear program as a valid benchmark and propose two LP-based parameterized online algorithms. Rigorous online competitive ratio analysis is offered to demonstrate the flexibility and efficiency of our online algorithms in balancing the two conflicting goals, promotions of fairness and profit. Experimental results on a real-world dataset are provided as well, which confirm our theoretical predictions.
With the popularity of the Internet, traditional offline resource allocation has evolved into a new form, called online resource allocation. It features the online arrivals of agents in the system and the real-time decision-making requirement upon the arrival of each online agent. Both offline and online resource allocation have wide applications in various real-world matching markets ranging from ridesharing to crowdsourcing. There are some emerging applications such as rebalancing in bike sharing and trip-vehicle dispatching in ridesharing, which involve a two-stage resource allocation process. The process consists of an offline phase and another sequential online phase, and both phases compete for the same set of resources. In this paper, we propose a unified model which incorporates both offline and online resource allocation into a single framework. Our model assumes non-uniform and known arrival distributions for online agents in the second online phase, which can be learned from historical data. We propose a parameterized linear programming (LP)-based algorithm, which is shown to be at most a constant factor of $1/4$ from the optimal. Experimental results on the real dataset show that our LP-based approaches outperform the LP-agnostic heuristics in terms of robustness and effectiveness.
Deep implicit field regression methods are effective for 3D reconstruction from single-view images. However, the impact of different sampling patterns on the reconstruction quality is not well-understood. In this work, we first study the effect of point set discrepancy on the network training. Based on Farthest Point Sampling algorithm, we propose a sampling scheme that theoretically encourages better generalization performance, and results in fast convergence for SGD-based optimization algorithms. Secondly, based on the reflective symmetry of an object, we propose a feature fusion method that alleviates issues due to self-occlusions which makes it difficult to utilize local image features. Our proposed system Ladybird is able to create high quality 3D object reconstructions from a single input image. We evaluate Ladybird on a large scale 3D dataset (ShapeNet) demonstrating highly competitive results in terms of Chamfer distance, Earth Mover's distance and Intersection Over Union (IoU).