Satellite image analysis has important implications for land use, urbanization, and ecosystem monitoring. Deep learning methods can facilitate the analysis of different satellite modalities, such as electro-optical (EO) and synthetic aperture radar (SAR) imagery, by supporting knowledge transfer between the modalities to compensate for individual shortcomings. Recent progress has shown how distributional alignment of neural network embeddings can produce powerful transfer learning models by employing a sliced Wasserstein distance (SWD) loss. We analyze how this method can be applied to Sentinel-1 and -2 satellite imagery and develop several extensions toward making it effective in practice. In an application to few-shot Local Climate Zone (LCZ) prediction, we show that these networks outperform multiple common baselines on datasets with a large number of classes. Further, we provide evidence that instance normalization can significantly stabilize the training process and that explicitly shaping the embedding space using supervised contrastive learning can lead to improved performance.

Reliable and efficient validation technologies are critical for the recent development of multi-vehicle cooperation and vehicle-road-cloud integration. In this paper, we introduce our miniature experimental platform, Mixed Cloud Control Testbed (MCCT), developed based on a new notion of Mixed Digital Twin (mixedDT). Combining Mixed Reality with Digital Twin, mixedDT integrates the virtual and physical spaces into a mixed one, where physical entities coexist and interact with virtual entities via their digital counterparts. Under the framework of mixedDT, MCCT contains three major experimental platforms in the physical, virtual and mixed spaces respectively, and provides a unified access for various human-machine interfaces and external devices such as driving simulators. A cloud unit, where the mixed experimental platform is deployed, is responsible for fusing multi-platform information and assigning control instructions, contributing to synchronous operation and real-time cross-platform interaction. Particularly, MCCT allows for multi-vehicle coordination composed of different multi-source vehicles (\eg, physical vehicles, virtual vehicles and human-driven vehicles). Validations on vehicle platooning demonstrate the flexibility and scalability of MCCT.

Solving the challenges of automatic machine translation of Building Automation System text metadata is a crucial first step in efficiently deploying smart building applications. The vocabulary used to describe building metadata appears small compared to general natural languages, but each term has multiple commonly used abbreviations. Conventional machine learning techniques are inefficient since they need to learn many different forms for the same word, and large amounts of data must be used to train these models. It is also difficult to apply standard techniques such as tokenisation since this commonly results in multiple output tags being associated with a single input token, something traditional sequence labelling models do not allow. Finite State Transducers can model sequence-to-sequence tasks where the input and output sequences are different lengths, and they can be combined with language models to ensure a valid output sequence is generated. We perform a preliminary analysis into the use of transducer-based language models to parse and normalise building point metadata.

Self-supervised learning (SSL) has been able to leverage unlabeled data to boost the performance of automatic speech recognition (ASR) models when we have access to only a small amount of transcribed speech data. However, this raises the question of which subset of the available unlabeled data should be selected for transcription. Our work investigates different unsupervised data selection techniques for fine-tuning the HuBERT model under a limited transcription budget. We investigate the impact of speaker diversity, gender bias, and topic diversity on the downstream ASR performance. We also devise two novel techniques for unsupervised data selection: pre-training loss based data selection and the perplexity of byte pair encoded clustered units (PBPE) and we show how these techniques compare to pure random data selection. Finally, we analyze the correlations between the inherent characteristics of the selected fine-tuning subsets as well as how these characteristics correlate with the resultant word error rate. We demonstrate the importance of token diversity, speaker diversity, and topic diversity in achieving the best performance in terms of WER.

Learning robust feature matching between the template and search area is crucial for 3D Siamese tracking. The core of Siamese feature matching is how to assign high feature similarity on the corresponding points between the template and search area for precise object localization. In this paper, we propose a novel point cloud registration-driven Siamese tracking framework, with the intuition that spatially aligned corresponding points (via 3D registration) tend to achieve consistent feature representations. Specifically, our method consists of two modules, including a tracking-specific nonlocal registration module and a registration-aided Sinkhorn template-feature aggregation module. The registration module targets at the precise spatial alignment between the template and search area. The tracking-specific spatial distance constraint is proposed to refine the cross-attention weights in the nonlocal module for discriminative feature learning. Then, we use the weighted SVD to compute the rigid transformation between the template and search area, and align them to achieve the desired spatially aligned corresponding points. For the feature aggregation model, we formulate the feature matching between the transformed template and search area as an optimal transport problem and utilize the Sinkhorn optimization to search for the outlier-robust matching solution. Also, a registration-aided spatial distance map is built to improve the matching robustness in indistinguishable regions (e.g., smooth surface). Finally, guided by the obtained feature matching map, we aggregate the target information from the template into the search area to construct the target-specific feature, which is then fed into a CenterPoint-like detection head for object localization. Extensive experiments on KITTI, NuScenes and Waymo datasets verify the effectiveness of our proposed method.

With the rise of artificial intelligence (AI) in recent years and the subsequent increase in complexity of the applied models, the growing demand in computational resources is starting to pose a significant challenge. The need for higher compute power is being met with increasingly more potent accelerator hardware as well as the use of large and powerful compute clusters. However, the gain in prediction accuracy from large models trained on distributed and accelerated systems ultimately comes at the price of a substantial increase in energy demand, and researchers have started questioning the environmental friendliness of such AI methods at scale. Consequently, awareness of energy efficiency plays an important role for AI model developers and hardware infrastructure operators likewise. The energy consumption of AI workloads depends both on the model implementation and the composition of the utilized hardware. Therefore, accurate measurements of the power draw of respective AI workflows on different types of compute nodes is key to algorithmic improvements and the design of future compute clusters and hardware. Towards this end, we present measurements of the energy consumption of two typical applications of deep learning models on different types of heterogeneous compute nodes. Our results indicate that 1. contrary to common approaches, deriving energy consumption directly from runtime is not accurate, but the consumption of the compute node needs to be considered regarding its composition; 2. neglecting accelerator hardware on mixed nodes results in overproportional inefficiency regarding energy consumption; 3. energy consumption of model training and inference should be considered separately - while training on GPUs outperforms all other node types regarding both runtime and energy consumption, inference on CPU nodes can be comparably efficient. One advantage of our approach is the fact that the information on energy consumption is available to all users of the supercomputer and not just those with administrator rights, enabling an easy transfer to other workloads alongside a raise in user-awareness of energy consumption.

The environmental sustainability of machine learning models is increasingly receiving attention, however mostly from academia. Here, the conversation tends to focus on the carbon footprint of language models which are not necessarily representative of the general machine learning field and not enough attention is paid to the operations phase of machine learning models. In addition, existing material on the topic of the environmental impact of machine learning puts too little emphasis on how the environmental impact can actually be estimated and reduced. This article is an attempt to address these issues and is written for practitioners and researchers alike who do hands-on machine learning. Although this post was written with machine learning in mind, some of the contents is also applicable to general software engineering. Before we begin, I want to emphasise that this post was not written to point fingers or moralise. The aim is simply to present information that you may or may not find relevant to your daily activities. All software -- from the apps that run on our phones to the data science pipelines that run in the cloud -- consume electricity and as long as not all our electricity is generated by renewable energy sources, electricity consumption will have a carbon footprint.

Parameter-efficient methods (like Prompt or Adapters) for adapting pre-trained language models to downstream tasks have been popular recently. However, hindrances still prevent these methods from reaching their full potential. For example, two significant challenges are few-shot adaptation and cross-task generalization ability. To tackle these issues, we propose a general framework to enhance the few-shot adaptation and cross-domain generalization ability of parameter-efficient methods. In our framework, we prime the self-supervised model for parameter-efficient methods to rapidly adapt to various downstream few-shot tasks. To evaluate the authentic generalization ability of these parameter-efficient methods, we conduct experiments on a few-shot cross-domain benchmark containing 160 diverse NLP tasks. The experiment result reveals that priming by tuning PLM only with extra training tasks leads to the best performance. Also, we perform a comprehensive analysis of various parameter-efficient methods under few-shot cross-domain scenarios.

We consider a model where a signal (discrete or continuous) is observed with an additive Gaussian noise process. The signal is issued from a linear combination of a finite but increasing number of translated features. The features are continuously parameterized by their location and depend on some scale parameter. First, we extend previous prediction results for off-the-grid estimators by taking into account here that the scale parameter may vary. The prediction bounds are analogous, but we improve the minimal distance between two consecutive features locations in order to achieve these bounds. Next, we propose a goodness-of-fit test for the model and give non-asymptotic upper bounds of the testing risk and of the minimax separation rate between two distinguishable signals. In particular, our test encompasses the signal detection framework. We deduce upper bounds on the minimal energy, expressed as the 2-norm of the linear coefficients, to successfully detect a signal in presence of noise. The general model considered in this paper is a non-linear extension of the classical high-dimensional regression model. It turns out that, in this framework, our upper bound on the minimax separation rate matches (up to a logarithmic factor) the lower bound on the minimax separation rate for signal detection in the high dimensional linear model associated to a fixed dictionary of features. We also propose a procedure to test whether the features of the observed signal belong to a given finite collection under the assumption that the linear coefficients may vary, but do not change to opposite signs under the null hypothesis. A non-asymptotic upper bound on the testing risk is given. We illustrate our results on the spikes deconvolution model with Gaussian features on the real line and with the Dirichlet kernel, frequently used in the compressed sensing literature, on the torus.

Recent advances in Federated Learning (FL) have paved the way towards the design of novel strategies for solving multiple learning tasks simultaneously, by leveraging cooperation among networked devices. Multi-Task Learning (MTL) exploits relevant commonalities across tasks to improve efficiency compared with traditional transfer learning approaches. By learning multiple tasks jointly, significant reduction in terms of energy footprints can be obtained. This article provides a first look into the energy costs of MTL processes driven by the Model-Agnostic Meta-Learning (MAML) paradigm and implemented in distributed wireless networks. The paper targets a clustered multi-task network setup where autonomous agents learn different but related tasks. The MTL process is carried out in two stages: the optimization of a meta-model that can be quickly adapted to learn new tasks, and a task-specific model adaptation stage where the learned meta-model is transferred to agents and tailored for a specific task. This work analyzes the main factors that influence the MTL energy balance by considering a multi-task Reinforcement Learning (RL) setup in a robotized environment. Results show that the MAML method can reduce the energy bill by at least 2 times compared with traditional approaches without inductive transfer. Moreover, it is shown that the optimal energy balance in wireless networks depends on uplink/downlink and sidelink communication efficiencies.