Search

Closed-loop simulation is essential for advancing end-to-end autonomous driving systems. Contemporary sensor simulation methods, such as NeRF and 3DGS, rely predominantly on conditions closely aligned with training data distributions, which are largely confined to forward-driving scenarios. Consequently, these methods face limitations when rendering complex maneuvers (e.g., lane change, acceleration, deceleration). Recent advancements in autonomous-driving world models have demonstrated the potential to generate diverse driving videos. However, these approaches remain constrained to 2D video generation, inherently lacking the spatiotemporal coherence required to capture intricacies of dynamic driving environments. In this paper, we introduce DriveDreamer4D, which enhances 4D driving scene representation leveraging world model priors. Specifically, we utilize the world model as a data machine to synthesize novel trajectory videos based on real-world driving data. Notably, we explicitly leverage structured conditions to control the spatial-temporal consistency of foreground and background elements, thus the generated data adheres closely to traffic constraints. To our knowledge, DriveDreamer4D is the first to utilize video generation models for improving 4D reconstruction in driving scenarios. Experimental results reveal that DriveDreamer4D significantly enhances generation quality under novel trajectory views, achieving a relative improvement in FID by 24.5%, 39.0%, and 10.5% compared to PVG, S3Gaussian, and Deformable-GS. Moreover, DriveDreamer4D markedly enhances the spatiotemporal coherence of driving agents, which is verified by a comprehensive user study and the relative increases of 20.3%, 42.0%, and 13.7% in the NTA-IoU metric., Comment: Project Page: https://drivedreamer4d.github.io

Recent works in volume rendering, \textit{e.g.} NeRF and 3D Gaussian Splatting (3DGS), significantly advance the rendering quality and efficiency with the help of the learned implicit neural radiance field or 3D Gaussians. Rendering on top of an explicit representation, the vanilla 3DGS and its variants deliver real-time efficiency by optimizing the parametric model with single-view supervision per iteration during training which is adopted from NeRF. Consequently, certain views are overfitted, leading to unsatisfying appearance in novel-view synthesis and imprecise 3D geometries. To solve aforementioned problems, we propose a new 3DGS optimization method embodying four key novel contributions: 1) We transform the conventional single-view training paradigm into a multi-view training strategy. With our proposed multi-view regulation, 3D Gaussian attributes are further optimized without overfitting certain training views. As a general solution, we improve the overall accuracy in a variety of scenarios and different Gaussian variants. 2) Inspired by the benefit introduced by additional views, we further propose a cross-intrinsic guidance scheme, leading to a coarse-to-fine training procedure concerning different resolutions. 3) Built on top of our multi-view regulated training, we further propose a cross-ray densification strategy, densifying more Gaussian kernels in the ray-intersect regions from a selection of views. 4) By further investigating the densification strategy, we found that the effect of densification should be enhanced when certain views are distinct dramatically. As a solution, we propose a novel multi-view augmented densification strategy, where 3D Gaussians are encouraged to get densified to a sufficient number accordingly, resulting in improved reconstruction accuracy., Comment: Project Page:https://xiaobiaodu.github.io/mvgs-project/

Training Deep Neural Networks (DNNs) with billions of parameters generally involves pipeline-parallel (PP) execution. Unfortunately, PP model training can use GPUs inefficiently, especially at large scale, due to idle GPU time caused by pipeline bubbles, which are often 15-30% and can exceed 60% of the training job's GPU allocation. To improve the GPU utilization of PP model training, this paper describes PipeFill, which fills pipeline bubbles with execution of other pending jobs. By leveraging bubble GPU time, PipeFill reduces the GPU utilization sacrifice associated with scaling-up of large-model training. To context-switch between fill jobs and the main training job with minimal overhead to the main job, and maximize fill job efficiency, PipeFill carefully fits fill job work to measured bubble durations and GPU memory availability, introduces explicit pipeline-bubble instructions, and orchestrates placement and execution of fill jobs in pipeline bubbles. Experiments show that PipeFill can increase overall utilization by up to 63% for GPUs used in large-scale LLM training, with <2% slowdown of the training job, and 5-15% even for low-scale LLM training. For large-scale LLM training on 8K GPUs, the 63% increase translates to up to 2.6K additional GPUs worth of work completed.

Large Language models have achieved impressive performance in automated software engineering. Extensive efforts have been made to evaluate the abilities of code LLMs in various aspects, with an increasing number of benchmarks and evaluation frameworks proposed. Apart from the most sought-after capability of code generation, the capability of code comprehension is being granted growing attention. Nevertheless, existing works assessing the code comprehension capability of LLMs exhibit varied limitations. Evaluation frameworks like CRUXEval and REval usually focus on code reasoning tasks over a certain input case, leading to a limited range of execution traces covered, resulting in a loss in code semantics examined and the inability to assess the comprehensive understanding of LLMs concerning the target program. To tackle the challenges above, we propose SpecEval, a novel black-box evaluation framework to evaluate code comprehension in LLMs via program specifications. Inspired by the idea that specifications can comprehensively articulate program behaviors concerning all possible execution traces, we employ formal specifications to represent program semantics and perform thorough evaluations. In particular, four specification-related tasks are designed to assess the capability of LLMs from basic to advanced levels. Moreover, counterfactual analysis is conducted to study the performance variance of LLMs under semantics-preserving perturbations, and progressive consistency analysis is performed to study the performance consistency of LLMs over a series of tasks with sequential dependence. Systematic experiments are conducted on six state-of-the-art LLMs. Experimental results present a below-satisfactory performance of LLMs on specification-related tasks, revealing the limitations of existing LLMs in articulating program semantics, underscoring future directions for enhancement.

Powerful foundation models, including large language models (LLMs), with Transformer architectures have ushered in a new era of Generative AI across various industries. Industry and research community have witnessed a large number of new applications, based on those foundation models. Such applications include question and answer, customer services, image and video generation, and code completions, among others. However, as the number of model parameters reaches to hundreds of billions, their deployment incurs prohibitive inference costs and high latency in real-world scenarios. As a result, the demand for cost-effective and fast inference using AI accelerators is ever more higher. To this end, our tutorial offers a comprehensive discussion on complementary inference optimization techniques using AI accelerators. Beginning with an overview of basic Transformer architectures and deep learning system frameworks, we deep dive into system optimization techniques for fast and memory-efficient attention computations and discuss how they can be implemented efficiently on AI accelerators. Next, we describe architectural elements that are key for fast transformer inference. Finally, we examine various model compression and fast decoding strategies in the same context., Comment: [v2] Tutorial website added [v1] Tutorial published at KDD 2024. Camera-ready version

Our objective is to leverage a differentiable radiance field \eg NeRF to reconstruct detailed 3D surfaces in addition to producing the standard novel view renderings. There have been related methods that perform such tasks, usually by utilizing a signed distance field (SDF). However, the state-of-the-art approaches still fail to correctly reconstruct the small-scale details, such as the leaves, ropes, and textile surfaces. Considering that different methods formulate and optimize the projection from SDF to radiance field with a globally constant Eikonal regularization, we improve with a ray-wise weighting factor to prioritize the rendering and zero-crossing surface fitting on top of establishing a perfect SDF. We propose to adaptively adjust the regularization on the signed distance field so that unsatisfying rendering rays won't enforce strong Eikonal regularization which is ineffective, and allow the gradients from regions with well-learned radiance to effectively back-propagated to the SDF. Consequently, balancing the two objectives in order to generate accurate and detailed surfaces. Additionally, concerning whether there is a geometric bias between the zero-crossing surface in SDF and rendering points in the radiance field, the projection becomes adjustable as well depending on different 3D locations during optimization. Our proposed \textit{RaNeuS} are extensively evaluated on both synthetic and real datasets, achieving state-of-the-art results on both novel view synthesis and geometric reconstruction., Comment: 3DV 2024, oral. In: Proceedings of the IEEE/CVF International Conference on 3D Vision (2023)

Diffusion models have emerged as dominant performers for image generation. To support training large diffusion models, this paper studies pipeline parallel training of diffusion models and proposes DiffusionPipe, a synchronous pipeline training system that advocates innovative pipeline bubble filling technique, catering to structural characteristics of diffusion models. State-of-the-art diffusion models typically include trainable (the backbone) and non-trainable (e.g., frozen input encoders) parts. We first unify optimal stage partitioning and pipeline scheduling of single and multiple backbones in representative diffusion models with a dynamic programming approach. We then propose to fill the computation of non-trainable model parts into idle periods of the pipeline training of the backbones by an efficient greedy algorithm, thus achieving high training throughput. Extensive experiments show that DiffusionPipe can achieve up to 1.41x speedup over pipeline parallel methods and 1.28x speedup over data parallel training on popular diffusion models.

The Mixture-of-Expert (MoE) technique plays a crucial role in expanding the size of DNN model parameters. However, it faces the challenge of extended all-to-all communication latency during the training process. Existing methods attempt to mitigate this issue by overlapping all-to-all with expert computation. Yet, these methods frequently fall short of achieving sufficient overlap, consequently restricting the potential for performance enhancements. In our study, we extend the scope of this challenge by considering overlap at the broader training graph level. During the forward pass, we enable non-MoE computations to overlap with all-to-all through careful partitioning and pipelining. In the backward pass, we achieve overlap with all-to-all by scheduling gradient weight computations. We implement these techniques in Lancet, a system using compiler-based optimization to automatically enhance MoE model training. Our extensive evaluation reveals that Lancet significantly reduces the time devoted to non-overlapping communication, by as much as 77%. Moreover, it achieves a notable end-to-end speedup of up to 1.3 times when compared to the state-of-the-art solutions., Comment: 11 pages, 16 figures. Published in MLSys'24

In surgical oncology, screening colonoscopy plays a pivotal role in providing diagnostic assistance, such as biopsy, and facilitating surgical navigation, particularly in polyp detection. Computer-assisted endoscopic surgery has recently gained attention and amalgamated various 3D computer vision techniques, including camera localization, depth estimation, surface reconstruction, etc. Neural Radiance Fields (NeRFs) and Neural Implicit Surfaces (NeuS) have emerged as promising methodologies for deriving accurate 3D surface models from sets of registered images, addressing the limitations of existing colon reconstruction approaches stemming from constrained camera movement. However, the inadequate tissue texture representation and confused scale problem in monocular colonoscopic image reconstruction still impede the progress of the final rendering results. In this paper, we introduce a novel method for colon section reconstruction by leveraging NeuS applied to endoscopic images, supplemented by a single frame of depth map. Notably, we pioneered the exploration of utilizing only one frame depth map in photorealistic reconstruction and neural rendering applications while this single depth map can be easily obtainable from other monocular depth estimation networks with an object scale. Through rigorous experimentation and validation on phantom imagery, our approach demonstrates exceptional accuracy in completely rendering colon sections, even capturing unseen portions of the surface. This breakthrough opens avenues for achieving stable and consistently scaled reconstructions, promising enhanced quality in cancer screening procedures and treatment interventions.

Getting large language models (LLMs) to perform well on the downstream tasks requires pre-training over trillions of tokens. This typically demands a large number of powerful computational devices in addition to a stable distributed training framework to accelerate the training. The growing number of applications leveraging AI/ML had led to a scarcity of the expensive conventional accelerators (such as GPUs), which begs the need for the alternative specialized-accelerators that are scalable and cost-efficient. AWS Trainium is the second-generation machine learning accelerator that has been purposely built for training large deep learning models. Its corresponding instance, Amazon EC2 trn1, is an alternative to GPU instances for LLM training. However, training LLMs with billions of parameters on trn1 is challenging due to its relatively nascent software ecosystem. In this paper, we showcase HLAT: a 7 billion parameter decoder-only LLM pre-trained using trn1 instances over 1.8 trillion tokens. The performance of HLAT is benchmarked against popular open source baseline models including LLaMA and OpenLLaMA, which have been trained on NVIDIA GPUs and Google TPUs, respectively. On various evaluation tasks, we show that HLAT achieves model quality on par with the baselines. We also share the best practice of using the Neuron Distributed Training Library (NDTL), a customized distributed training library for AWS Trainium to achieve efficient training. Our work demonstrates that AWS Trainium powered by the NDTL is able to successfully pre-train state-of-the-art LLM models with high performance and cost-effectiveness.

With the ongoing development of deep learning, an increasing number of AI models have surpassed the performance levels of human clinical practitioners. However, the prevalence of AI diagnostic products in actual clinical practice remains significantly lower than desired. One crucial reason for this gap is the so-called `black box' nature of AI models. Clinicians' distrust of black box models has directly hindered the clinical deployment of AI products. To address this challenge, we propose ContrastDiagnosis, a straightforward yet effective interpretable diagnosis framework. This framework is designed to introduce inherent transparency and provide extensive post-hoc explainability for deep learning model, making them more suitable for clinical medical diagnosis. ContrastDiagnosis incorporates a contrastive learning mechanism to provide a case-based reasoning diagnostic rationale, enhancing the model's transparency and also offers post-hoc interpretability by highlighting similar areas. High diagnostic accuracy was achieved with AUC of 0.977 while maintain a high transparency and explainability.

Multi-task model training has been adopted to enable a single deep neural network model (often a large language model) to handle multiple tasks (e.g., question answering and text summarization). Multi-task training commonly receives input sequences of highly different lengths due to the diverse contexts of different tasks. Padding (to the same sequence length) or packing (short examples into long sequences of the same length) is usually adopted to prepare input samples for model training, which is nonetheless not space or computation efficient. This paper proposes a dynamic micro-batching approach to tackle sequence length variation and enable efficient multi-task model training. We advocate pipeline-parallel training of the large model with variable-length micro-batches, each of which potentially comprises a different number of samples. We optimize micro-batch construction using a dynamic programming-based approach, and handle micro-batch execution time variation through dynamic pipeline and communication scheduling, enabling highly efficient pipeline training. Extensive evaluation on the FLANv2 dataset demonstrates up to 4.39x higher training throughput when training T5, and 3.25x when training GPT, as compared with packing-based baselines. DynaPipe's source code is publicly available at https://github.com/awslabs/optimizing-multitask-training-through-dynamic-pipelines., Comment: 18 pages, 18 figures

Serving deep learning (DL) models on relational data has become a critical requirement across diverse commercial and scientific domains, sparking growing interest recently. In this visionary paper, we embark on a comprehensive exploration of representative architectures to address the requirement. We highlight three pivotal paradigms: The state-of-the-art DL-centric architecture offloads DL computations to dedicated DL frameworks. The potential UDF-centric architecture encapsulates one or more tensor computations into User Defined Functions (UDFs) within the relational database management system (RDBMS). The potential relation-centric architecture aims to represent a large-scale tensor computation through relational operators. While each of these architectures demonstrates promise in specific use scenarios, we identify urgent requirements for seamless integration of these architectures and the middle ground in-between these architectures. We delve into the gaps that impede the integration and explore innovative strategies to close them. We present a pathway to establish a novel RDBMS for enabling a broad class of data-intensive DL inference applications., Comment: * Authors are ordered alphabetically; Jia Zou is the corresponding author

An important challenge in Machine Learning compilers like XLA is multi-pass optimization and analysis. There has been recent interest chiefly in XLA target-dependent optimization on the graph-level, subgraph-level, and kernel-level phases. We specifically focus on target-independent optimization XLA HLO pass ordering: our approach aims at finding the optimal sequence of compiler optimization passes, which is decoupled from target-dependent optimization. However, there is little domain specific study in pass ordering for XLA HLO. To this end, we propose introducing deep Reinforcement Learning (RL) based search for optimal XLA HLO pass ordering. We also propose enhancements to the deep RL algorithms to further improve optimal search performance and open the research direction for domain-specific guidance for RL. We create an XLA Gym experimentation framework as a tool to enable RL algorithms to interact with the compiler for passing optimizations and thereby train agents. Overall, in our experimentation we observe an average of $13.3\%$ improvement in operation count reduction on a benchmark of GPT-2 training graphs and $10.4\%$ improvement on a diverse benchmark including GPT-2, BERT, and ResNet graphs using the proposed approach over the compiler's default phase ordering., Comment: Workshop on ML for Systems @ NeurIPS 2022

Recently, large vision model, Segment Anything Model (SAM), has revolutionized the computer vision field, especially for image segmentation. SAM presented a new promptable segmentation paradigm that exhibit its remarkable zero-shot generalization ability. An extensive researches have explore the potential and limits of SAM in various downstream tasks. In this study, we presents $\mathrm{SAM^{Med}}$, an enhanced framework for medical image annotation that leverages the capabilities of SAM. $\mathrm{SAM^{Med}}$ framework consisted of two submodules, namely $\mathrm{SAM^{assist}}$ and $\mathrm{SAM^{auto}}$. The $\mathrm{SAM^{assist}}$ demonstrates the generalization ability of SAM to the downstream medical segmentation task using the prompt-learning approach. Results show a significant improvement in segmentation accuracy with only approximately 5 input points. The $\mathrm{SAM^{auto}}$ model aims to accelerate the annotation process by automatically generating input prompts. The proposed SAP-Net model achieves superior segmentation performance with only five annotated slices, achieving an average Dice coefficient of 0.80 and 0.82 for kidney and liver segmentation, respectively. Overall, $\mathrm{SAM^{Med}}$ demonstrates promising results in medical image annotation. These findings highlight the potential of leveraging large-scale vision models in medical image annotation tasks.

As deep learning is pervasive in modern applications, many deep learning frameworks are presented for deep learning practitioners to develop and train DNN models rapidly. Meanwhile, as training large deep learning models becomes a trend in recent years, the training throughput and memory footprint are getting crucial. Accordingly, optimizing training workloads with compiler optimizations is inevitable and getting more and more attentions. However, existing deep learning compilers (DLCs) mainly target inference and do not incorporate holistic optimizations, such as automatic differentiation and automatic mixed precision, in training workloads. In this paper, we present RAF, a deep learning compiler for training. Unlike existing DLCs, RAF accepts a forward model and in-house generates a training graph. Accordingly, RAF is able to systematically consolidate graph optimizations for performance, memory and distributed training. In addition, to catch up to the state-of-the-art performance with hand-crafted kernel libraries as well as tensor compilers, RAF proposes an operator dialect mechanism to seamlessly integrate all possible kernel implementations. We demonstrate that by in-house training graph generation and operator dialect mechanism, we are able to perform holistic optimizations and achieve either better training throughput or larger batch size against PyTorch (eager and torchscript mode), XLA, and DeepSpeed for popular transformer models on GPUs.

Recent years have seen an increase in the development of large deep learning (DL) models, which makes training efficiency crucial. Common practice is struggling with the trade-off between usability and performance. On one hand, DL frameworks such as PyTorch use dynamic graphs to facilitate model developers at a price of sub-optimal model training performance. On the other hand, practitioners propose various approaches to improving the training efficiency by sacrificing some of the flexibility, ranging from making the graph static for more thorough optimization (e.g., XLA) to customizing optimization towards large-scale distributed training (e.g., DeepSpeed and Megatron-LM). In this paper, we aim to address the tension between usability and training efficiency through separation of concerns. Inspired by DL compilers that decouple the platform-specific optimizations of a tensor-level operator from its arithmetic definition, this paper proposes a schedule language, Slapo, to decouple model execution from definition. Specifically, Slapo works on a PyTorch model and uses a set of schedule primitives to convert the model for common model training optimizations such as high-performance kernels, effective 3D parallelism, and efficient activation checkpointing. Compared to existing optimization solutions, Slapo progressively optimizes the model "as-needed" through high-level primitives, and thus preserving programmability and debuggability for users to a large extent. Our evaluation results show that by scheduling the existing hand-crafted optimizations in a systematic way using Slapo, we are able to improve training throughput by up to 2.92x on a single machine with 8 NVIDIA V100 GPUs, and by up to 1.41x on multiple machines with up to 64 GPUs, when compared to the out-of-the-box performance of DeepSpeed and Megatron-LM., Comment: Accepted to ASPLOS'24

Lidar became an important component of the perception systems in autonomous driving. But challenges of training data acquisition and annotation made emphasized the role of the sensor to sensor domain adaptation. In this work, we address the problem of lidar upsampling. Learning on lidar point clouds is rather a challenging task due to their irregular and sparse structure. Here we propose a method for lidar point cloud upsampling which can reconstruct fine-grained lidar scan patterns. The key idea is to utilize edge-aware dense convolutions for both feature extraction and feature expansion. Additionally applying a more accurate Sliced Wasserstein Distance facilitates learning of the fine lidar sweep structures. This in turn enables our method to employ a one-stage upsampling paradigm without the need for coarse and fine reconstruction. We conduct several experiments to evaluate our method and demonstrate that it provides better upsampling.

As deep learning models nowadays are widely adopted by both cloud services and edge devices, reducing the latency of deep learning model inferences becomes crucial to provide efficient model serving. However, it is challenging to develop efficient tensor programs for deep learning operators due to the high complexity of modern accelerators and the rapidly growing number of operators. Deep learning compilers, such as Apache TVM, adopt declarative scheduling primitives to lower the bar of developing tensor programs. However, we show that this approach is insufficient to cover state-of-the-art tensor program optimizations. In this paper, we propose to embed the scheduling process into tensor programs and use dedicated mappings, called task mappings, to define the computation assignment and ordering. This new approach greatly enriches the expressible optimizations by allowing developers to manipulate tensor programs at a much finer granularity. We call the proposed method the task-mapping programming paradigm. In addition, we propose a new post-scheduling fusion optimization that allows developers to focus on scheduling every single operator and automates the fusion after scheduling. It greatly reduces the engineering efforts for operator fusion. Our proposed paradigm also constructs an efficient hardware-centric schedule space, which is agnostic to the program input size and greatly reduces the tuning time. With the proposed paradigm, we implement a deep learning compiler Hidet. Extensive experiments on modern convolution and transformer models show that Hidet outperforms state-of-the-art DNN inference framework, ONNX Runtime, and compiler, TVM equipped with scheduler AutoTVM and Ansor, by up to 1.48x (1.22x on average). It also reduces the tuning time by 20x and 11x compared with AutoTVM and Ansor, respectively. We open-sourced hidet at https://www.github.com/hidet-org/hidet., Comment: 15 pages, 22 figures, 1 table

We propose a novel convolutional operator for the task of point cloud completion. One striking characteristic of our approach is that, conversely to related work it does not require any max-pooling or voxelization operation. Instead, the proposed operator used to learn the point cloud embedding in the encoder extracts permutation-invariant features from the point cloud via a soft-pooling of feature activations, which are able to preserve fine-grained geometric details. These features are then passed on to a decoder architecture. Due to the compression in the encoder, a typical limitation of this type of architectures is that they tend to lose parts of the input shape structure. We propose to overcome this limitation by using skip connections specifically devised for point clouds, where links between corresponding layers in the encoder and the decoder are established. As part of these connections, we introduce a transformation matrix that projects the features from the encoder to the decoder and vice-versa. The quantitative and qualitative results on the task of object completion from partial scans on the ShapeNet dataset show that incorporating our approach achieves state-of-the-art performance in shape completion both at low and high resolutions., Comment: Accepted in International Journal of Computer Vision

Existing general purpose frameworks for gigantic model training, i.e., dense models with billions of parameters, cannot scale efficiently on cloud environment with various networking conditions due to large communication overheads. In this paper, we propose MiCS, which Minimizes the Communication Scale to bring down communication overhead. Specifically, by decreasing the number of participants in a communication collective, MiCS can utilize heterogeneous network bandwidth, reduce network traffic over slower links, reduce the latency of communications for maintaining high network bandwidth utilization, and amortize expensive global gradient synchronization overhead. Our evaluation on AWS shows that the system throughput of MiCS is up to 2.89$\times$ that of the state-of-the-art large model training systems. MiCS achieves near-linear scaling efficiency, which is up to 1.27$\times$ that of DeepSpeed. MiCS allows us to train a proprietary model with 100 billion parameters on 512 GPUs with 99.4% weak-scaling efficiency, and it is able to saturate over 54.5% theoretical computation power of each GPU on a public cloud with less GPU memory and more restricted networks than DGX-A100 clusters.

Lung cancer has the highest mortality rate of deadly cancers in the world. Early detection is essential to treatment of lung cancer. However, detection and accurate diagnosis of pulmonary nodules depend heavily on the experiences of radiologists and can be a heavy workload for them. Computer-aided diagnosis (CAD) systems have been developed to assist radiologists in nodule detection and diagnosis, greatly easing the workload while increasing diagnosis accuracy. Recent development of deep learning, greatly improved the performance of CAD systems. However, lack of model reliability and interpretability remains a major obstacle for its large-scale clinical application. In this work, we proposed a multi-task explainable deep-learning model for pulmonary nodule diagnosis. Our neural model can not only predict lesion malignancy but also identify relevant manifestations. Further, the location of each manifestation can also be visualized for visual interpretability. Our proposed neural model achieved a test AUC of 0.992 on LIDC public dataset and a test AUC of 0.923 on our in-house dataset. Moreover, our experimental results proved that by incorporating manifestation identification tasks into the multi-task model, the accuracy of the malignancy classification can also be improved. This multi-task explainable model may provide a scheme for better interaction with the radiologists in a clinical environment.

We propose a novel approach aimed at object and semantic scene completion from a partial scan represented as a 3D point cloud. Our architecture relies on three novel layers that are used successively within an encoder-decoder structure and specifically developed for the task at hand. The first one carries out feature extraction by matching the point features to a set of pre-trained local descriptors. Then, to avoid losing individual descriptors as part of standard operations such as max-pooling, we propose an alternative neighbor-pooling operation that relies on adopting the feature vectors with the highest activations. Finally, up-sampling in the decoder modifies our feature extraction in order to increase the output dimension. While this model is already able to achieve competitive results with the state of the art, we further propose a way to increase the versatility of our approach to process point clouds. To this aim, we introduce a second model that assembles our layers within a transformer architecture. We evaluate both architectures on object and indoor scene completion tasks, achieving state-of-the-art performance., Comment: Conference on Computer Vision and Pattern Recognition (CVPR) 2022

There have been numerous recently proposed methods for monocular depth prediction (MDP) coupled with the equally rapid evolution of benchmarking tools. However, we argue that MDP is currently witnessing benchmark over-fitting and relying on metrics that are only partially helpful to gauge the usefulness of the predictions for 3D applications. This limits the design and development of novel methods that are truly aware of - and improving towards estimating - the 3D structure of the scene rather than optimizing 2D-based distances. In this work, we aim to bring structural awareness to MDP, an inherently 3D task, by exhibiting the limits of evaluation metrics towards assessing the quality of the 3D geometry. We propose a set of metrics well suited to evaluate the 3D geometry of MDP approaches and a novel indoor benchmark, RIO-D3D, crucial for the proposed evaluation methodology. Our benchmark is based on a real-world dataset featuring high-quality rendered depth maps obtained from RGB-D reconstructions. We further demonstrate this to help benchmark the closely-tied task of 3D scene completion.

Alpa automates model-parallel training of large deep learning (DL) models by generating execution plans that unify data, operator, and pipeline parallelism. Existing model-parallel training systems either require users to manually create a parallelization plan or automatically generate one from a limited space of model parallelism configurations. They do not suffice to scale out complex DL models on distributed compute devices. Alpa distributes the training of large DL models by viewing parallelisms as two hierarchical levels: inter-operator and intra-operator parallelisms. Based on it, Alpa constructs a new hierarchical space for massive model-parallel execution plans. Alpa designs a number of compilation passes to automatically derive efficient parallel execution plans at each parallelism level. Alpa implements an efficient runtime to orchestrate the two-level parallel execution on distributed compute devices. Our evaluation shows Alpa generates parallelization plans that match or outperform hand-tuned model-parallel training systems even on models they are designed for. Unlike specialized systems, Alpa also generalizes to models with heterogeneous architectures and models without manually-designed plans. Alpa's source code is publicly available at https://github.com/alpa-projects/alpa, Comment: OSDI 2022

Although pre-trained language models have remarkably enhanced the generation ability of dialogue systems, open-domain Chinese dialogue systems are still limited by the dialogue data and the model size compared with English ones. In this paper, we propose EVA, a Chinese dialogue system that contains the largest Chinese pre-trained dialogue model with 2.8B parameters. To build this model, we collect the largest Chinese dialogue dataset named WDC-Dialogue from various public social media. This dataset contains 1.4B context-response pairs and is used as the pre-training corpus of EVA. Extensive experiments on automatic and human evaluation show that EVA outperforms other Chinese pre-trained dialogue models especially in the multi-turn interaction of human-bot conversations., Comment: 8 pages, 4 figures

Most modern deep learning-based multi-view 3D reconstruction techniques use RNNs or fusion modules to combine information from multiple images after independently encoding them. These two separate steps have loose connections and do not allow easy information sharing among views. We propose LegoFormer, a transformer model for voxel-based 3D reconstruction that uses the attention layers to share information among views during all computational stages. Moreover, instead of predicting each voxel independently, we propose to parametrize the output with a series of low-rank decomposition factors. This reformulation allows the prediction of an object as a set of independent regular structures then aggregated to obtain the final reconstruction. Experiments conducted on ShapeNet demonstrate the competitive performance of our model with respect to the state of the art while having increased interpretability thanks to the self-attention layers. We also show promising generalization results to real data.

This paper propose to combine pretrained language models with the modular dialogue paradigm for open-domain dialogue modeling. Our method, semantic-enhanced finetuning, instantiates conversation understanding, planning, and response generation as a language model finetuning task. At inference, we disentangle semantic and token variations by specifying sampling methods and constraints for each module separately. For training and evaluation, we present X-Weibo, a Chinese multi-turn open-domain dialogue dataset with automatic annotation for emotions, DAs, and topical words. Experiments show that semantic-enhanced finetuning outperforms strong baselines on non-semantic and semantic metrics, improves the human-evaluated relevance, coherence, and informativeness, and exhibits considerable controllability over semantic variables., Comment: Under review

Neural dialogue generation models trained with the one-hot target distribution suffer from the over-confidence issue, which leads to poor generation diversity as widely reported in the literature. Although existing approaches such as label smoothing can alleviate this issue, they fail to adapt to diverse dialog contexts. In this paper, we propose an Adaptive Label Smoothing (AdaLabel) approach that can adaptively estimate a target label distribution at each time step for different contexts. The maximum probability in the predicted distribution is used to modify the soft target distribution produced by a novel light-weight bi-directional decoder module. The resulting target distribution is aware of both previous and future contexts and is adjusted to avoid over-training the dialogue model. Our model can be trained in an end-to-end manner. Extensive experiments on two benchmark datasets show that our approach outperforms various competitive baselines in producing diverse responses., Comment: ACL2021 Main Track (Long Paper), Code Available in https://github.com/lemon234071/AdaLabel

Deep neural networks (DNNs) have been ubiquitously applied in many applications, and accelerators are emerged as an enabler to support the fast and efficient inference tasks of these applications. However, to achieve high model coverage with high performance, each accelerator vendor has to develop a full compiler stack to ingest, optimize, and execute the DNNs. This poses significant challenges in the development and maintenance of the software stack. In addition, the vendors have to contiguously update their hardware and/or software to cope with the rapid evolution of the DNN model architectures and operators. To address these issues, this paper proposes an open source framework that enables users to only concentrate on the development of their proprietary code generation tools by reusing as many as possible components in the existing deep learning compilers. Our framework provides users flexible and easy-to-use interfaces to partition their models into segments that can be executed on "the best" processors to take advantage of the powerful computation capability of accelerators. Our case study shows that our framework has been deployed in multiple commercial vendors' compiler stacks with only a few thousand lines of code.

In this paper, we investigate the optimal design of a wireless-powered covert communication (WP-CC) system, in which a probabilistic accumulate-then-transmit (ATT) protocol is proposed to maximize the communication covertness subject to a quality-of-service (QoS) requirement on communication. Specifically, in the considered WP-CC system, a full-duplex (FD) receiver transmits artificial noise (AN) to simultaneously charge an energy-constrained transmitter and to confuse a warden's detection on the transmitter's communication activity. With the probabilistic ATT protocol, the transmitter sends its information with a prior probability, i.e., $p$, conditioned on the available energy being sufficient. Our analysis shows that the probabilistic ATT protocol can achieve higher covertness than the traditional ATT protocol with $p=1$. In order to facilitate the optimal design of the WP-CC system, we also derive the warden's minimum detection error probability and characterize the effective covert rate from the transmitter to the receiver to quantify the communication covertness and quality, respectively. The derived analytical results facilitate the joint optimization of the probability $p$ and the information transmit power. We further present the optimal design of a cable-powered covert communication (CP-CC) system as a benchmark for comparison. Our simulation shows that the proposed probabilistic ATT protocol (with a varying $p$) can achieve the covertness upper bound determined by the CP-CC system, while the traditional ATT protocol (with $p=1$) cannot, which again confirms the benefits brought by the proposed probabilistic ATT in covert communications., Comment: 12 pages, 6 figures

Nowadays the world is facing a prominent paradox regarding thermal energy. The production of heat accounts for more than 50% of global final energy consumption while the waste heat potential analysis reveals that 72% of the global primary energy consumption is lost after conversion mainly in the form of heat. Towards global decarbonization, it is of vital importance to establish a solution to thermal energy utilization under full control. Here, we propose and realize an unprecedented concept -- barocaloric thermal batteries based on the inverse colossal barocaloric effect of NH4SCN. Thermal charging is initialized upon pressurization through an order-to-disorder phase transition below 364 K and in turn the discharging of 43 J g-1, which are eleven times more than the input mechanical energy, occurs on demand at depressurization at lower temperatures. The discharging is also manifested as a directly measured temperature rise of 12 K. The thermodynamic equilibrium nature of the pressure-restrained heat-carrying phase guarantees stable storage and/or transport over a variety of temporal and/or spatial scales. The barocaloric thermal batteries reinforced by their solid microscopic mechanism are expected to significantly advance the ability to take advantage of waste heat., Comment: 5 figures in the main text plus the Extended Data. Adiabatic-temperature-change data were added

Because of the increasing demand for computation in DNN, researchers develope both hardware and software mechanisms to reduce the compute and memory burden. A widely adopted approach is to use mixed precision data types. However, it is hard to leverage mixed precision without hardware support because of the overhead of data casting. Hardware vendors offer tensorized instructions for mixed-precision tensor operations, like Intel VNNI, Tensor Core, and ARM-DOT. These instructions involve a computing idiom that reduces multiple low precision elements into one high precision element. The lack of compilation techniques for this makes it hard to utilize these instructions: Using vendor-provided libraries for computationally-intensive kernels is inflexible and prevents further optimizations, and manually writing hardware intrinsics is error-prone and difficult for programmers. Some prior works address this problem by creating compilers for each instruction. This requires excessive effort when it comes to many tensorized instructions. In this work, we develop a compiler framework to unify the compilation for these instructions -- a unified semantics abstraction eases the integration of new instructions, and reuses the analysis and transformations. Tensorized instructions from different platforms can be compiled via UNIT with moderate effort for favorable performance. Given a tensorized instruction and a tensor operation, UNIT automatically detects the applicability, transforms the loop organization of the operation,and rewrites the loop body to leverage the tensorized instruction. According to our evaluation, UNIT can target various mainstream hardware platforms. The generated end-to-end inference model achieves 1.3x speedup over Intel oneDNN on an x86 CPU, 1.75x speedup over Nvidia cuDNN on an NvidiaGPU, and 1.13x speedup over a carefully tuned TVM solution for ARM DOT on an ARM CPU., Comment: 13 pages, 13 figures, and 1 table

Current low-precision quantization algorithms often have the hidden cost of conversion back and forth from floating point to quantized integer values. This hidden cost limits the latency improvement realized by quantizing Neural Networks. To address this, we present HAWQV3, a novel mixed-precision integer-only quantization framework. The contributions of HAWQV3 are the following: (i) An integer-only inference where the entire computational graph is performed only with integer multiplication, addition, and bit shifting, without any floating point operations or even integer division; (ii) A novel hardware-aware mixed-precision quantization method where the bit-precision is calculated by solving an integer linear programming problem that balances the trade-off between model perturbation and other constraints, e.g., memory footprint and latency; (iii) Direct hardware deployment and open source contribution for 4-bit uniform/mixed-precision quantization in TVM, achieving an average speed up of $1.45\times$ for uniform 4-bit, as compared to uniform 8-bit for ResNet50 on T4 GPUs; and (iv) extensive evaluation of the proposed methods on ResNet18/50 and InceptionV3, for various model compression levels with/without mixed precision. For ResNet50, our INT8 quantization achieves an accuracy of $77.58\%$, which is $2.68\%$ higher than prior integer-only work, and our mixed-precision INT4/8 quantization can reduce INT8 latency by $23\%$ and still achieve $76.73\%$ accuracy. Our framework and the TVM implementation have been open sourced.

Graph neural networks (GNNs) are gaining increasing popularity as a promising approach to machine learning on graphs. Unlike traditional graph workloads where each vertex/edge is associated with a scalar, GNNs attach a feature tensor to each vertex/edge. This additional feature dimension, along with consequently more complex vertex- and edge-wise computations, has enormous implications on locality and parallelism, which existing graph processing systems fail to exploit. This paper proposes FeatGraph to accelerate GNN workloads by co-optimizing graph traversal and feature dimension computation. FeatGraph provides a flexible programming interface to express diverse GNN models by composing coarse-grained sparse templates with fine-grained user-defined functions (UDFs) on each vertex/edge. FeatGraph incorporates optimizations for graph traversal into the sparse templates and allows users to specify optimizations for UDFs with a feature dimension schedule (FDS). FeatGraph speeds up end-to-end GNN training and inference by up to 32x on CPU and 7x on GPU., Comment: SC'20; changed all figures to type 1

Point clouds are often the default choice for many applications as they exhibit more flexibility and efficiency than volumetric data. Nevertheless, their unorganized nature -- points are stored in an unordered way -- makes them less suited to be processed by deep learning pipelines. In this paper, we propose a method for 3D object completion and classification based on point clouds. We introduce a new way of organizing the extracted features based on their activations, which we name soft pooling. For the decoder stage, we propose regional convolutions, a novel operator aimed at maximizing the global activation entropy. Furthermore, inspired by the local refining procedure in Point Completion Network (PCN), we also propose a patch-deforming operation to simulate deconvolutional operations for point clouds. This paper proves that our regional activation can be incorporated in many point cloud architectures like AtlasNet and PCN, leading to better performance for geometric completion. We evaluate our approach on different 3D tasks such as object completion and classification, achieving state-of-the-art accuracy., Comment: accepted in ECCV 2020 as oral