LG - 机器学习 CV - 计算机视觉 CL - 计算与语言 AS - 音频与语音 RO - 机器人

转自爱可可爱生活

1、[LG] SMORE: Knowledge Graph Completion and Multi-hop Reasoning in Massive Knowledge Graphs

H Ren, H Dai, B Dai, X Chen, D Zhou, J Leskovec, D Schuurmans

[Stanford University & Google Brain & UC Berkeley.]

SMORE: 海量知识图谱的知识图谱补全和多跳推理。知识图谱(KG)以头-关系-尾三元组的形式捕获知识，是许多人工智能系统的重要组成部分。在知识图谱上有两个重要的推理任务：（1）单跳知识图谱补全，涉及到知识图谱中单个链接的预测；（2）多跳推理，目标是预测哪些知识图谱实体满足给定的逻辑查询。基于嵌入的方法通过首先计算每个实体和关系的嵌入来解决这两个任务，然后用它们来形成预测。然而，现有的可扩展的KG嵌入框架只支持单跳知识图谱补全，不能应用于更具挑战性的多跳推理任务。本文提出可扩展多跳推理(SMORE)，第一个在KG中同时进行单跳和多跳推理的通用框架。只用一台机器，SMORE可以在Freebase KG(86M实体，338M边)中进行多跳推理，比以前考虑的KG大1500倍。SMORE运行时性能的关键是一种新的双向拒绝采样，实现了在线训练数据生成复杂性的平方根减少。此外，SMORE利用了异步调度、基于CPU重叠数据采样、基于GPU嵌入计算以及频繁CPU-GPU IO。SMORE以最小的GPU内存需求(86M节点Freebase上训练400维嵌入的2GB内存)将吞吐量(即训练速度)比之前的多跳KG框架提高了2.2倍，且随着GPU数量的增加实现了接近线性的加速。此外，在更简单的单跳知识图谱完成任务上，SMORE在单GPU和多GPU设置上都实现了与最先进的框架相当甚至更好的运行时性能。

Knowledge graphs (KGs) capture knowledge in the form of head–relation–tail triples and are a crucial component in many AI systems. There are two important reasoning tasks on KGs: (1) single-hop knowledge graph completion, which involves predicting individual links in the KG; and (2), multi-hop reasoning, where the goal is to predict which KG entities satisfy a given logical query. Embedding-based methods solve both tasks by first computing an embedding for each entity and relation, then using them to form predictions. However, existing scalable KG embedding frameworks only support single-hop knowledge graph completion and cannot be applied to the more challenging multi-hop reasoning task. Here we present Scalable Multi-hOp REasoning (SMORE), the first general framework for both single-hop and multi-hop reasoning in KGs. Using a single machine SMORE can perform multi-hop reasoning in Freebase KG (86M entities, 338M edges), which is 1,500× larger than previously considered KGs. The key to SMORE’s runtime performance is a novel bidirectional rejection sampling that achieves a square root reduction of the complexity of online training data generation. Furthermore, SMORE exploits asynchronous scheduling, overlapping CPU-based data sampling, GPU-based embedding computation, and frequent CPU–GPU IO. SMORE increases throughput (i.e., training speed) over prior multi-hop KG frameworks by 2.2× with minimal GPU memory requirements (2GB for training 400-dim embeddings on 86M-node Freebase) and achieves near linear speed-up with the number of GPUs. Moreover, on the simpler single-hop knowledge graph completion task SMORE achieves comparable or even better runtime performance to state-of-the-art frameworks on both single GPU and multi-GPU settings.

https://weibo.com/1402400261/L0tiR0Uwc

2、[LG] Identifying Nonlinear Dynamical Systems From Multi-Modal Time Series Data

P L Bommer, D Kramer, C Tombolini, G Koppe, D Durstewitz

[Heidelberg University]

多模态时序数据的非线性动态系统识别。在物理学、生物学或医学领域，根据经验观察到的时间序列通常是由一些潜在的动态系统(DS)产生的，这些动态系统是科学关注的目标。人们越来越有兴趣用机器学习的方法，以一种完全数据驱动的、无监督的方式重建这个潜在的动态系统。在许多科学领域，从许多数据模式中同时采样观察时间序列是很常见的，例如，在典型的神经科学实验中，电生理和行为时间序列。然而，目前用于重建动态系统的机器学习工具通常只关注一种数据模式。本文提出一种用于多模态数据整合的通用框架，以实现非线性动态系统识别和跨模态预测的。该框架基于动态可解释的递归神经网络，作为非线性动态系统的通用近似器，与广义线性模型类的特定模态解码器模型组相耦合。用于模型训练的期望最大化算法和变分推理算法都得到了提升和比较。在非线性DS基准上表明，该算法可通过利用其他渠道有效补偿一个数据渠道的过于嘈杂或缺失的信息，并在实验性神经科学数据上证明算法如何学习将不同的数据域与基本动态联系起来。

Empirically observed time series in physics, biology, or medicine, are commonly generated by some underlying dynamical system (DS) which is the target of scientific interest. There is an increasing interest to harvest machine learning methods to reconstruct this latent DS in a completely data-driven, unsupervised way. In many areas of science it is common to sample time series observations from many data modalities simultaneously, e.g. electrophysiological and behavioral time series in a typical neuroscience experiment. However, current machine learning tools for reconstructing DSs usually focus on just one data modality. Here we propose a general framework for multi-modal data integration for the purpose of nonlinear DS identification and cross-modal prediction. This framework is based on dynamically interpretable recurrent neural networks as general approximators of nonlinear DSs, coupled to sets of modality-specific decoder models from the class of generalized linear models. Both an expectation-maximization and a variational inference algorithm for model training are advanced and compared. We show on nonlinear DS benchmarks that our algorithms can efficiently compensate for too noisy or missing information in one data channel by exploiting other channels, and demonstrate on experimental neuroscience data how the algorithm learns to link different data domains to the underlying dynamics.

https://weibo.com/1402400261/L0tnbzGgO

3、[LG] Large Language Models Can Be Strong Differentially Private Learners

X Li, F Tramèr, P Liang, T Hashimoto

[Stanford University & Google Research]

大规模语言模型作为强大的差分隐私学习器。差分隐私(DP)学习在构建大规模文本深度学习模型方面取得了有限的成功，而将差分隐私随机梯度下降(DP-SGD)直接应用于NLP任务的尝试导致了巨大的性能下降和高计算开销。本文表明，这种性能下降可以通过以下方式得到缓解：（1）采用大规模预训练模型；（2）适合差分隐私优化的超参数；以及（3）与预训练程序一致的微调目标。在这些因素的作用下，获得了超越最先进的隐私训练方法和强大的非隐私基线的隐私NLP模型——直接在中等规模的语料库上用差分隐私优化对预训练模型进行微调。为解决在运行DP-SGD时遇到的计算挑战，提出一种节省内存的技术，允许DP-SGD中的剪切运行，而不需要为模型中的任何一层实例化梯度。该技术使隐私训练Transformers的内存成本与非隐私训练的内存成本几乎相同，而且运行时间开销不大。与传统观点相反的是，差分隐私优化在学习高维模型时失败了(由于噪声随维度的变化而变化)，经验结果显示，用预训练的模型进行隐私学习往往不会受到维度相关的性能下降的影响。

Differentially Private (DP) learning has seen limited success for building large deep learning models of text, and attempts at straightforwardly applying Differentially Private Stochastic Gradient Descent (DP-SGD) to NLP tasks have resulted in large performance drops and high computational overhead. We show that this performance drop can be mitigated with (1) the use of large pretrained models; (2) hyperparameters that suit DP optimization; and (3) fine-tuning objectives aligned with the pretraining procedure. With these factors set right, we obtain private NLP models that outperform state-of-the-art private training approaches and strong nonprivate baselines—by directly fine-tuning pretrained models with DP optimization on moderately-sized corpora. To address the computational challenge of running DP-SGD with large Transformers, we propose a memory saving technique that allows clipping in DP-SGD to run without instantiating per-example gradients for any layer in the model. The technique enables privately training Transformers with almost the same memory cost as non-private training at a modest run-time overhead. Contrary to conventional wisdom that DP optimization fails at learning high-dimensional models (due to noise that scales with dimension) empirical results reveal that private learning with pretrained models tends to not suffer from dimension-dependent performance degradation.

https://weibo.com/1402400261/L0tqEuCDn

4、[LG] Bounds all around: training energy-based models with bidirectional bounds

C Geng, J Wang, Z Gao, J Frellsen, S Hauberg

[Shanghai Jiao Tong University & Technical University of Denmark]

Bounds all around：基于双向界的基于能量模型训练。基于能量的模型(EBM)为密度估计提供了一个优雅的框架，但它们是出了名的难以训练。最近的工作建立了与生成对抗网络的联系，其中EBM通过一个具有变分价值函数的极小极大博弈来训练。本文提出了对EBM对数似然的双向约束，即在解决最小化博弈时，最大化下限并最小化上限。将一个界与稳定训练的梯度惩罚联系起来，从而为最佳工程实践提供基础。为了评估这些界，本文提出一种新的、高效的EBM发生器雅各比决定式的估计器。这些发展大大稳定了训练，并产生高质量的密度估计和样本生成。

Energy-based models (EBMs) provide an elegant framework for density estimation, but they are notoriously difficult to train. Recent work has established links to generative adversarial networks, where the EBM is trained through a minimax game with a variational value function. We propose a bidirectional bound on the EBM log-likelihood, such that we maximize a lower bound and minimize an upper bound when solving the minimax game. We link one bound to a gradient penalty that stabilizes training, thereby providing grounding for best engineering practice. To evaluate the bounds we develop a new and efficient estimator of the Jacobi-determinant of the EBM generator. We demonstrate that these developments significantly stabilize training and yield high-quality density estimation and sample generation.

https://weibo.com/1402400261/L0tuh5foO

5、[LG] Adversarial Intrinsic Motivation for Reinforcement Learning

I Durugkar, M Tec, S Niekum, P Stone

[The University of Texas at Austin]

强化学习对抗性内在动机。以最小化与参考分布的不匹配为目标的学习已被证明对生成建模和模仿学习是有用的。本文研究了这样一个目标，即策略的状态访问分布和目标分布之间的Wasserstein-1距离，是否可以有效地用于强化学习(RL)任务。本文关注的是目标条件下的强化学习，其中理想化的(无法实现的)目标分布在目标上有充分的措施。本文介绍了马尔科夫决策过程(MDP)特有的类比法，并使用该类比法来估计上述Wasserstein-1距离。进一步表明，最小化这个Wasserstein-1距离的策略是在尽可能少的步骤中达到目标的策略。所提出的方法，称为对抗性内在动机(AIM)，通过其双重目标估计这个Wasserstein-1距离，并用它来计算一个补充奖励函数。实验表明，该奖励函数随着MDP中的转换而平滑变化，并引导智能体的探索有效地找到目标。此外，将对抗性内在动机与事后经验重放(HER)结合起来，并表明与其他鼓励探索或加速学习的奖励相比，所产生的算法在几个模拟的机器人任务上大大加速了学习。

Learning with an objective to minimize the mismatch with a reference distribution has been shown to be useful for generative modeling and imitation learning. In this paper, we investigate whether one such objective, the Wasserstein-1 distance between a policy’s state visitation distribution and a target distribution, can be utilized effectively for reinforcement learning (RL) tasks. Specifically, this paper focuses on goal-conditioned reinforcement learning where the idealized (unachievable) target distribution has full measure at the goal. This paper introduces a quasimetric specific to Markov Decision Processes (MDPs) and uses this quasimetric to estimate the above Wasserstein-1 distance. It further shows that the policy that minimizes this Wasserstein-1 distance is the policy that reaches the goal in as few steps as possible. Our approach, termed Adversarial Intrinsic Motivation (AIM), estimates this Wasserstein-1 distance through its dual objective and uses it to compute a supplemental reward function. Our experiments show that this reward function changes smoothly with respect to transitions in the MDP and directs the agent’s exploration to find the goal efficiently. Additionally, we combine AIM with Hindsight Experience Replay (HER) and show that the resulting algorithm accelerates learning significantly on several simulated robotics tasks when compared to other rewards that encourage exploration or accelerate learning.

https://weibo.com/1402400261/L0txYzkbk

另外几篇值得关注的论文：

 

[CV] PP-PicoDet: A Better Real-Time Object Detector on Mobile Devices

PP-PicoDet：面向移动端更好的实时目标检测器

G Yu, Q Chang, W Lv, C Xu, C Cui, W Ji, Q Dang, K Deng, G Wang, Y Du, B Lai, Q Liu, X Hu, D Yu, Y Ma

[Baidu]

https://weibo.com/1402400261/L0tDCr8F3

[CV] Rebooting ACGAN: Auxiliary Classifier GANs with Stable Training

重启ACGAN：稳定训练辅助分类器GAN

M Kang, W Shim, M Cho, J Park

[Pohang University of Science and Technology (POSTECH)]

https://weibo.com/1402400261/L0tFHlGs9

[CV] DIB-R++: Learning to Predict Lighting and Material with a Hybrid Differentiable Renderer

DIB-R++：基于混合可微渲染器的光照和材料预测

W Chen, J Litalien, J Gao, Z Wang, C F Tsang, S Khamis, O Litany, S Fidler

[NVIDIA & McGill University]

https://weibo.com/1402400261/L0tID13S7

[CV] DeepVecFont: Synthesizing High-quality Vector Fonts via Dual-modality Learning

DeepVecFont：基于双模态学习的高质量矢量字体合成

Y Wang, Z Lian

[Peking University]

https://weibo.com/1402400261/L0tLjpVnB

内容中包含的图片若涉及版权问题，请及时与我们联系删除