**Bi3D** is a stereo depth estimation framework that estimates depth via a series of binary classifications. Rather than testing if objects are at a particular depth *D*, as existing stereo methods do, it classifies them as being closer or farther than *D*. It takes the stereo pair and a disparity $d\_{i}$ and produces a confidence map, which can be thresholded to yield the binary segmentation. To estimate depth on $N + 1$ quantization levels we run this network $N$ times and maximize the probability in Equation 8 (see paper). To estimate continuous depth, whether full or selective, we run the [SegNet](https://paperswithcode.com/method/segnet) block of Bi3DNet for each disparity level and work directly on the confidence volume.

Bi3D

Bi3D: Stereo Depth Estimation via Binary Classifications

Nlogistic-sigmoid function (NLSIG) is a modern logistic-sigmoid function definition for modelling growth (or decay) processes. It features two logistic metrics (YIR and XIR) for monitoring growth from a two-dimensional (x-y axis) perspective.

NLSIG

The **Cross-Attention** module is an attention module used in [CrossViT](https://paperswithcode.com/method/crossvit) for fusion of multi-scale features. The CLS token of the large branch (circle) serves as a query token to interact with the patch tokens from the small branch through attention. $f\left(·\right)$ and $g\left(·\right)$ are projections to align dimensions. The small branch follows the same procedure but swaps CLS and patch tokens from another branch.

Cross-Attention Module

CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification

**DeepCluster** is a self-supervision approach for learning image representations.  DeepCluster iteratively groups the features with a standard clustering algorithm, k-means, and uses the subsequent assignments as supervision to update
the weights of the network

DeepCluster

Deep Clustering for Unsupervised Learning of Visual Features

**Meta-augmentation** helps generate more varied tasks for a single example in meta-learning. It can be distinguished from data augmentation in classic machine learning as follows. For data augmentation in classical machine learning, the aim is to generate more varied examples, within a single task. Meta-augmentation has the exact opposite aim: we wish to generate more varied tasks,
for a single example, to force the learner to quickly learn a new task from feedback. In meta-augmentation, adding randomness discourages the base learner and model from learning trivial solutions that do not generalize to new tasks.

Meta-augmentation

Meta-Learning Requires Meta-Augmentation

**AutoDropout** automates the process of designing [dropout](https://paperswithcode.com/method/dropout) patterns using a [Transformer](https://paperswithcode.com/method/transformer) based controller. In this method, a controller learns to generate a dropout pattern at every channel and layer of a target network, such as a [ConvNet](https://paperswithcode.com/methods/category/convolutional-neural-networks) or a Transformer. The target network is then trained with the dropped-out pattern, and its resulting validation performance is used as a signal for the controller to learn from. The resulting pattern is applied to a convolutional output channel, which is a common building block of image recognition models.

The controller network generates the tokens to describe the configurations of the dropout pattern. The tokens are generated like words in a language model. For every layer in a ConvNet, a group of 8 tokens need to be made to create a dropout pattern. These 8 tokens are generated sequentially. In the figure above, size, stride, and repeat indicate the size and the tiling of the pattern; rotate, shear_x, and shear_y specify the geometric transformations of the pattern; share_c is a binary deciding whether a pattern is applied to all $C$ channels; and residual is a binary deciding whether the pattern is applied to the residual branch as well. If we need $L$ dropout patterns, the controller will generate $8L$ decisions.

AutoDropout

AutoDropout: Learning Dropout Patterns to Regularize Deep Networks

This method applies Polya-Gamma latent variables as a way to obtain closed form expressions for full-conditionals of posterior distributions in sampling algorithms like MCMC.

Polya-Gamma Augmentation

Bayesian inference for logistic models using Polya-Gamma latent variables

**NEAT**, or **Neural Attention Fields**, is a feature representation for end-to-end imitation learning models. NEAT is a continuous function which maps locations in Bird's Eye View (BEV) scene coordinates to waypoints and semantics, using intermediate attention maps to iteratively compress high-dimensional 2D image features into a compact representation. This allows the model to selectively attend to relevant regions in the input while ignoring information irrelevant to the driving task, effectively associating the images with the BEV representation. Furthermore, visualizing the attention maps for models with NEAT intermediate representations provides improved interpretability.

NEAT

NEAT: Neural Attention Fields for End-to-End Autonomous Driving

A variant of [CutMix](https://paperswithcode.com/method/cutmix) which randomly samples masks from Fourier space.

FMix

FMix: Enhancing Mixed Sample Data Augmentation

Spatio-temporal features extraction that measure the stabilty. The proposed method is based on a compression algorithm named Run Length Encoding. The workflow of the method is presented bellow.

**ByteScheduler** is a generic communication scheduler for distributed DNN training acceleration. It is based on analysis that partitioning and rearranging the tensor transmissions can result in optimal results in theory and good performance in real-world even with scheduling overhead.

ByteScheduler

**Parallax** is a hybrid parallel method for training large neural networks. Parallax is a framework that optimizes data parallel training by utilizing the sparsity of model parameters. Parallax introduces a hybrid approach that combines Parameter Server and AllReduce architectures to optimize the amount of data transfer according to the sparsity.

Parallax pursues a hybrid approach that uses the Parameter Server architecture for handling sparse variables and the AllReduce architecture for handling dense variables. Moreover, Parallax partitions large sparse variables by a near-optimal number of partitions to maximize parallelism while maintaining low computation and communication overhead. Parallax further optimizes training with local aggregation and smart operation placement to mitigate communication overhead. Graph transformation in Parallax automatically applies all of these optimizations and the data parallel training itself at the framework level to minimize user efforts for writing and optimizing a distributed program by composing low-level primitives.

Parallax

Please enter a description about the method here

VocGAN

VocGAN: A High-Fidelity Real-time Vocoder with a Hierarchically-nested Adversarial Network

**VIME **, or **Value Imputation and Mask Estimation**, is a self- and semi-supervised learning framework for tabular data. It consists of a pretext task of estimating mask vectors from corrupted tabular data in addition to the reconstruction pretext task for self-supervised learning.

VIME

VIME: Extending the Success of Self- and Semi-supervised Learning to Tabular Domain

A **CNN BiLSTM** is a hybrid bidirectional [LSTM](https://paperswithcode.com/method/lstm) and CNN architecture. In the original formulation applied to named entity recognition, it learns both character-level and word-level features. The CNN component is used to induce the character-level features. For each word the model employs a [convolution](https://paperswithcode.com/method/convolution) and a [max pooling](https://paperswithcode.com/method/max-pooling) layer to extract a new feature vector from the per-character feature vectors such as character embeddings and (optionally) character type.

CNN BiLSTM

Named Entity Recognition with Bidirectional LSTM-CNNs

**Enhanced Sequential Inference Model** or **ESIM** is a sequential NLI model proposed in [Enhanced LSTM for Natural Language Inference](https://www.aclweb.org/anthology/P17-1152) paper.

ESIM

Enhanced LSTM for Natural Language Inference

**YOLOP** is a panoptic driving perception network for handling traffic object detection, drivable area segmentation and lane detection simultaneously. It is composed of one encoder for feature extraction and three decoders to handle the specific tasks. It can be thought of a lightweight version of Tesla's HydraNet model for self-driving cars.

A lightweight CNN, from Scaled-yolov4, is used as the encoder to extract features from the image. Then these feature maps are fed to three decoders to complete their respective tasks. The detection decoder is based on the current best-performing single-stage detection network, [YOLOv4](https://paperswithcode.com/method/yolov4),  for two main reasons: (1) The single-stage detection network is faster than the two-stage detection network. (2) The grid-based prediction mechanism of the single-stage detector is more related to the other two semantic segmentation tasks, while instance segmentation is usually combined with the region based detector as in [Mask R-CNN](https://paperswithcode.com/method/mask-r-cnn). The feature map output by the encoder incorporates semantic features of different levels and scales, and our segmentation branch can use these feature maps to complete pixel-wise semantic prediction.

YOLOP

YOLOP: You Only Look Once for Panoptic Driving Perception

A [GCN](https://paperswithcode.com/method/gcn) method targeted at the unique spectral properties of dense graphs and hypergraphs, enabled by efficient numerical linear algebra.

Viet-Anh on Software

AI Glossary

Tech Terminology

Bi3D

nlogistic-sigmoid function

Cross-Attention Module

DeepCluster

Meta-augmentation

AutoDropout

Data augmentation using Polya-Gamma latent variables.

Neural Attention Fields

FMix

Spatio-temporal stability analysis

ByteScheduler

Parallax

VocGAN

Value Imputation and Mask Estimation

CNN Bidirectional LSTM

Enhanced Sequential Inference Model

YOLOP

Pseudoinverse Graph Convolutional Network

Viet-Anh on Software