Pre-processing for whole slide images can affect classification performance both in the training and inference stages. Our study analyzes the impact of pre-processing parameters on inference and training across single- and multiple-domain datasets. However, searching for an optimal parameter set is time-consuming. To overcome this, we propose a novel Similarity-based Simulated Annealing approach for fast parameter tuning to enhance inference performance on single-domain data. Our method demonstrates significant performance improvements in accuracy, which raise accuracy from 0.512 to 0.847 in a single domain. We further extend our insight into training performance in multi-domain data by employing a novel Bayesian optimization to search optimal pre-processing parameters, resulting in a high AUC of 0.967. We highlight that better pre-processing for WSI can contribute to further accuracy improvement in the histology area.
Retrieval-based augmentations that aim to incorporate knowledge from an external database into language models have achieved great success in various knowledge-intensive (KI) tasks, such as question-answering and text generation. However, integrating retrievals in non-knowledge-intensive (NKI) tasks, such as text classification, is still challenging. Existing works focus on concatenating retrievals to inputs as context to form the prompt-based inputs. Unfortunately, such methods require language models to have the capability to handle long texts. Besides, inferring such concatenated data would also consume a significant amount of computational resources. To solve these challenges, we propose \textbf{ReFusion} in this paper, a computation-efficient \textbf{Re}trieval representation \textbf{Fusion} with neural architecture search. The main idea is to directly fuse the retrieval representations into the language models. Specifically, we first propose an online retrieval module that retrieves representations of similar sentences. Then, we present a retrieval fusion module including two effective ranking schemes, i.e., reranker-based scheme and ordered-mask-based scheme, to fuse the retrieval representations with hidden states. Furthermore, we use Neural Architecture Search (NAS) to seek the optimal fusion structure across different layers. Finally, we conduct comprehensive experiments, and the results demonstrate our ReFusion can achieve superior and robust performance on various NKI tasks.
Recent works on learned index open a new direction for the indexing field. The key insight of the learned index is to approximate the mapping between keys and positions with piece-wise linear functions. Such methods require partitioning key space for a better approximation. Although lots of heuristics are proposed to improve the approximation quality, the bottleneck is that the segmentation overheads could hinder the overall performance. This paper tackles the approximation problem by applying a \textit{distribution transformation} to the keys before constructing the learned index. A two-stage Normalizing-Flow-based Learned index framework (NFL) is proposed, which first transforms the original complex key distribution into a near-uniform distribution, then builds a learned index leveraging the transformed keys. For effective distribution transformation, we propose a Numerical Normalizing Flow (Numerical NF). Based on the characteristics of the transformed keys, we propose a robust After-Flow Learned Index (AFLI). To validate the performance, comprehensive evaluations are conducted on both synthetic and real-world workloads, which shows that the proposed NFL produces the highest throughput and the lowest tail latency compared to the state-of-the-art learned indexes.
Deep-learning-based compressor has received interests recently due to much improved compression ratio. However, modern approaches suffer from long execution time. To ease this problem, this paper targets on cutting down the execution time of deep-learning-based compressors. Building history-dependencies sequentially (e.g., recurrent neural networks) is responsible for long inference latency. Instead, we introduce transformer into deep learning compressors to build history-dependencies in parallel. However, existing transformer is too heavy in computation and incompatible to compression tasks. This paper proposes a fast general-purpose lossless compressor, TRACE, by designing a compression-friendly structure based on a single-layer transformer. We first design a new metric to advise the selection part of compression model structures. Byte-grouping and Shared-ffn schemes are further proposed to fully utilize the capacity of the single-layer transformer. These features allow TRACE to achieve competitive compression ratio and a much faster speed. In addition, we further accelerate the compression procedure by designing a controller to reduce the parameter updating overhead. Experiments show that TRACE achieves an overall $\sim$3x speedup while keeps a comparable compression ratio to the state-of-the-art compressors. The source code for TRACE and links to the datasets are available at https://github.com/mynotwo/A-Fast-Transformer-based-General-Purpose-LosslessCompressor.
Deep neural networks with batch normalization (BN-DNNs) are invariant to weight rescaling due to their normalization operations. However, using weight decay (WD) benefits these weight-scale-invariant networks, which is often attributed to an increase of the effective learning rate when the weight norms are decreased. In this paper, we demonstrate the insufficiency of the previous explanation and investigate the implicit biases of stochastic gradient descent (SGD) on BN-DNNs to provide a theoretical explanation for the efficacy of weight decay. We identity two implicit biases of SGD on BN-DNNs: 1) the weight norms in SGD training remain constant in the continuous-time domain and keep increasing in the discrete-time domain; 2) SGD optimizes weight vectors in fully-connected networks or convolution kernels in convolution neural networks by updating components lying in the input feature span, while leaving those components orthogonal to the input feature span unchanged. Thus, SGD without WD accumulates weight noise orthogonal to the input feature span, and cannot eliminate such noise. Our empirical studies corroborate the hypothesis that weight decay suppresses weight noise that is left untouched by SGD. Furthermore, we propose to use weight rescaling (WRS) instead of weight decay to achieve the same regularization effect, while avoiding performance degradation of WD on some momentum-based optimizers. Our empirical results on image recognition show that regardless of optimization methods and network architectures, training BN-DNNs using WRS achieves similar or better performance compared with using WD. We also show that training with WRS generalizes better compared to WD, on other computer vision tasks.
Nested networks or slimmable networks are neural networks whose architectures can be adjusted instantly during testing time, e.g., based on computational constraints. Recent studies have focused on a "nested dropout" layer, which is able to order the nodes of a layer by importance during training, thus generating a nested set of sub-networks that are optimal for different configurations of resources. However, the dropout rate is fixed as a hyper-parameter over different layers during the whole training process. Therefore, when nodes are removed, the performance decays in a human-specified trajectory rather than in a trajectory learned from data. Another drawback is the generated sub-networks are deterministic networks without well-calibrated uncertainty. To address these two problems, we develop a Bayesian approach to nested neural networks. We propose a variational ordering unit that draws samples for nested dropout at a low cost, from a proposed Downhill distribution, which provides useful gradients to the parameters of nested dropout. Based on this approach, we design a Bayesian nested neural network that learns the order knowledge of the node distributions. In experiments, we show that the proposed approach outperforms the nested network in terms of accuracy, calibration, and out-of-domain detection in classification tasks. It also outperforms the related approach on uncertainty-critical tasks in computer vision.
Using weight decay to penalize the L2 norms of weights in neural networks has been a standard training practice to regularize the complexity of networks. In this paper, we show that a family of regularizers, including weight decay, is ineffective at penalizing the intrinsic norms of weights for networks with positively homogeneous activation functions, such as linear, ReLU and max-pooling functions. As a result of homogeneity, functions specified by the networks are invariant to the shifting of weight scales between layers. The ineffective regularizers are sensitive to such shifting and thus poorly regularize the model capacity, leading to overfitting. To address this shortcoming, we propose an improved regularizer that is invariant to weight scale shifting and thus effectively constrains the intrinsic norm of a neural network. The derived regularizer is an upper bound for the input gradient of the network so minimizing the improved regularizer also benefits the adversarial robustness. Residual connections are also considered and we show that our regularizer also forms an upper bound to input gradients of such a residual network. We demonstrate the efficacy of our proposed regularizer on various datasets and neural network architectures at improving generalization and adversarial robustness.
Estimating the uncertainty of a Bayesian model has been investigated for decades. The model posterior is almost always intractable, such that approximation is necessary. In many real-world cases, even though a decent estimation of the model posterior is obtained, another approximation is required to compute the predictive distribution over the desired output. A common accurate solution is to use Monte Carlo (MC) integration. However, it needs to maintain a large number of samples, evaluate the model repeatedly and average multiple model outputs. In this paper, we propose a method to approximate the probability distribution over the simplex induced by model posterior, enabling tractable computation of the predictive distribution for classification. The aim is to approximate the induced uncertainty of a specific Bayesian model, meanwhile alleviating the heavy workload of MC integration in testing time. Methodologically, we adapt Wasserstein distance to learn the induced conditional distributions, which is novel for Bayesian learning. The proposed method is universally applicable to Bayesian classification models that allow for posterior sampling. Empirical results validate the strong practical performance of our approach.