Real-world applications with multiple sensors observing an event are expected
to make continuously-available predictions, even in cases where information may
be intermittently missing. We explore methods in ensemble learning and sensor
fusion to make use of redundancy and information shared between four camera
views, applied to the task of hand activity classification for autonomous
driving. In particular, we show that a late-fusion approach between parallel
convolutional neural networks can outperform even the best-placed single camera
model. To enable this approach, we propose a scheme for handling missing
information, and then provide comparative analysis of this late-fusion approach
to additional methods such as weighted majority voting and model combination
schemes.

本文探讨了多传感器融合和集成学习方法在自动驾驶的手部活动识别任务中的应用，展示了采用并行卷积神经网络作为传感器间信息融合的一种迟到融合方法，提出了处理丢失信息的方案，并对该方法进行了比较分析。

缺失数据下的多视角集成学习：计算框架和使用 Safe Autonomous Driving 领域新数据的评估

Multi-View Ensemble Learning With Missing Data: Computational Framework and Evaluations using Novel Data from the Safe Autonomous Driving Domain

We investigate a late-fusion approach to piano transcription, combined with a
strong temporal prior in the form of a handcrafted Hidden Markov Model (HMM).
The network architecture under consideration is compact in terms of its number
of parameters and easy to train with gradient descent. The network outputs are
fused over time in the final stage to obtain note segmentations, with an HMM
whose transition probabilities are chosen based on a model of attack, decay,
sustain, release (ADSR) envelopes, commonly used for sound synthesis. The note
segments are then subject to a final binary decision rule to reject too weak
note segment hypotheses. We obtain state-of-the-art results on the MAPS
dataset, and are able to outperform other approaches by a large margin, when
predicting complete note regions from onsets to offsets.

研究了一种迟到融合方法和手工隐藏马尔可夫模型在钢琴转录中的应用，利用 ADSR 包络获取转移概率进行音符分段并经过最终的二进制决策规则，从而实现对完整音符的准确转换。在 MAPS 数据集上取得了最先进的结果，并能够大幅超越其他方法。