连续序列到序列建模的分层状态空间模型

Hierarchical State Space Models for Continuous Sequence-to-Sequence Modeling

February 15, 2024

作者: Raunaq Bhirangi, Chenyu Wang, Venkatesh Pattabiraman, Carmel Majidi, Abhinav Gupta, Tess Hellebrekers, Lerrel Pinto

cs.AI

摘要

从原始感官数据序列推理是一个普遍存在的问题，涵盖领域从医疗设备到机器人技术。这些问题通常涉及使用长序列的原始传感器数据（例如磁力计、压阻器）来预测理想物理量的序列（例如力量、惯性测量）。虽然传统方法对于局部线性预测问题很有效，但在使用真实世界传感器时往往效果不佳。这些传感器通常是非线性的，受到外部变量（例如振动）的影响，并且表现出数据相关漂移。对于许多问题，由于获取基准标签需要昂贵设备，预测任务会因标记数据集较小而变得更加困难。在这项工作中，我们提出了分层状态空间模型（HiSS），这是一种概念简单、新颖的连续序列预测技术。HiSS将结构化状态空间模型堆叠在一起，形成时间层次结构。在从基于触觉的状态预测到基于加速度计的惯性测量等六个真实世界传感器数据集中，HiSS在均方误差上至少比现有的序列模型（如因果Transformer、LSTM、S4和Mamba）表现出至少23%的优越性。我们的实验进一步表明，HiSS在小型数据集上表现出高效的扩展性，并且与现有的数据过滤技术兼容。代码、数据集和视频可在https://hiss-csp.github.io找到。

English

Reasoning from sequences of raw sensory data is a ubiquitous problem across fields ranging from medical devices to robotics. These problems often involve using long sequences of raw sensor data (e.g. magnetometers, piezoresistors) to predict sequences of desirable physical quantities (e.g. force, inertial measurements). While classical approaches are powerful for locally-linear prediction problems, they often fall short when using real-world sensors. These sensors are typically non-linear, are affected by extraneous variables (e.g. vibration), and exhibit data-dependent drift. For many problems, the prediction task is exacerbated by small labeled datasets since obtaining ground-truth labels requires expensive equipment. In this work, we present Hierarchical State-Space Models (HiSS), a conceptually simple, new technique for continuous sequential prediction. HiSS stacks structured state-space models on top of each other to create a temporal hierarchy. Across six real-world sensor datasets, from tactile-based state prediction to accelerometer-based inertial measurement, HiSS outperforms state-of-the-art sequence models such as causal Transformers, LSTMs, S4, and Mamba by at least 23% on MSE. Our experiments further indicate that HiSS demonstrates efficient scaling to smaller datasets and is compatible with existing data-filtering techniques. Code, datasets and videos can be found on https://hiss-csp.github.io.