Determining the optimal cost function parameters of Model Predictive Control (MPC) to optimize multiple control objectives is a challenging and time-consuming task. Multiobjective Bayesian Optimization (BO) techniques solve this problem by determining a Pareto optimal parameter set for an MPC with static weights. However, a single parameter set may not deliver the most optimal closed-loop control performance when the context of the MPC operating conditions changes during its operation, urging the need to adapt the cost function weights at runtime. Deep Reinforcement Learning (RL) algorithms can automatically learn context-dependent optimal parameter sets and dynamically adapt for a Weightsvarying MPC (WMPC). However, learning cost function weights from scratch in a continuous action space may lead to unsafe operating states. To solve this, we propose a novel approach limiting the RL actions within a safe learning space representing a catalog of pre-optimized BO Pareto-optimal weight sets. We conceive a RL agent not to learn in a continuous space but to proactively anticipate upcoming control tasks and to choose the most optimal discrete actions, each corresponding to a single set of Pareto optimal weights, context-dependent. Hence, even an untrained RL agent guarantees a safe and optimal performance. Experimental results demonstrate that an untrained RL-WMPC shows Pareto-optimal closed-loop behavior and training the RL-WMPC helps exhibit a performance beyond the Pareto-front.

Nonlinear hyperspectral unmixing has recently received considerable attention, as linear mixture models do not lead to an acceptable resolution in some problems. In fact, most nonlinear unmixing methods are designed by assuming specific assumptions on the nonlinearity model which subsequently limits the unmixing performance. In this paper, we propose an unsupervised nonlinear unmixing approach based on deep learning by incorporating a general nonlinear model with no special assumptions. This model consists of two branches. In the first branch, endmembers are learned by reconstructing the rows of hyperspectral images using some hidden layers, and in the second branch, abundance values are learned based on the columns of respective images. Then, using multi-task learning, we introduce an auxiliary task to enforce the two branches to work together. This technique can be considered as a regularizer mitigating overfitting, which improves the performance of the total network. Extensive experiments on synthetic and real data verify the effectiveness of the proposed method compared to some state-of-the-art hyperspectral unmixing methods.

Recent advancements in data-driven approaches, such as Neural Operator (NO), have demonstrated their effectiveness in reducing the solving time of Partial Differential Equations (PDEs). However, one major challenge faced by these approaches is the requirement for a large amount of high-precision training data, which needs significant computational costs during the generation process. To address this challenge, we propose a novel PDE dataset generation algorithm, namely Differential Operator Action in Solution space (DiffOAS), which speeds up the data generation process and enhances the precision of the generated data simultaneously. Specifically, DiffOAS obtains a few basic PDE solutions and then combines them to get solutions. It applies differential operators on these solutions, a process we call 'operator action', to efficiently generate precise PDE data points. Theoretical analysis shows that the time complexity of DiffOAS method is one order lower than the existing generation method. Experimental results show that DiffOAS accelerates the generation of large-scale datasets with 10,000 instances by 300 times. Even with just 5% of the generation time, NO trained on the data generated by DiffOAS exhibits comparable performance to that using the existing generation method, which highlights the efficiency of DiffOAS.

Tree ensemble algorithms as RandomForest and GradientBoosting are currently the dominant methods for modeling discrete or tabular data, however, they are unable to perform a hierarchical representation learning from raw data as NeuralNetworks does thanks to its multi-layered structure, which is a key feature for DeepLearning problems and modeling unstructured data. This limitation is due to the fact that tree algorithms can not be trained with back-propagation because of their mathematical nature. However, in this work, we demonstrate that the mathematical formulation of bagging and boosting can be combined together to define a graph-structured-tree-ensemble algorithm with a distributed representation learning process between trees naturally (without using back-propagation). We call this novel approach Distributed Gradient Boosting Forest (DGBF) and we demonstrate that both RandomForest and GradientBoosting can be expressed as particular graph architectures of DGBT. Finally, we see that the distributed learning outperforms both RandomForest and GradientBoosting in 7 out of 9 datasets.

Air pollution constitutes a global problem of paramount importance that affects not only human health, but also the environment. The existence of spatial and temporal data regarding the concentrations of pollutants is crucial for performing air pollution studies and monitor emissions. However, although observation data presents great temporal coverage, the number of stations is very limited and they are usually built in more populated areas. The main objective of this work is to create models capable of inferring pollutant concentrations in locations where no observation data exists. A machine learning model, more specifically the random forest model, was developed for predicting concentrations in the Iberian Peninsula in 2019 for five selected pollutants: $NO_2$, $O_3$ $SO_2$, $PM10$, and $PM2.5$. Model features include satellite measurements, meteorological variables, land use classification, temporal variables (month, day of year), and spatial variables (latitude, longitude, altitude). The models were evaluated using various methods, including station 10-fold cross-validation, in which in each fold observations from 10\% of the stations are used as testing data and the rest as training data. The $R^2$, RMSE and mean bias were determined for each model. The $NO_2$ and $O_3$ models presented good values of $R^2$, 0.5524 and 0.7462, respectively. However, the $SO_2$, $PM10$, and $PM2.5$ models performed very poorly in this regard, with $R^2$ values of -0.0231, 0.3722, and 0.3303, respectively. All models slightly overestimated the ground concentrations, except the $O_3$ model. All models presented acceptable cross-validation RMSE, except the $O_3$ and $PM10$ models where the mean value was a little higher (12.5934 $\mu g/m^3$ and 10.4737 $\mu g/m^3$, respectively).

Embedded camera systems are ubiquitous, representing the most widely deployed example of a wireless embedded system. They capture a representation of the world - the surroundings illuminated by visible or infrared light. Despite their widespread usage, the architecture of embedded camera systems has remained unchanged, which leads to limitations. They visualize only a tiny portion of the world. Additionally, they are energy-intensive, leading to limited battery lifespan. We present PixelGen, which re-imagines embedded camera systems. Specifically, PixelGen combines sensors, transceivers, and low-resolution image and infrared vision sensors to capture a broader world representation. They are deliberately chosen for their simplicity, low bitrate, and power consumption, culminating in an energy-efficient platform. We show that despite the simplicity, the captured data can be processed using transformer-based image and language models to generate novel representations of the environment. For example, we demonstrate that it can allow the generation of high-definition images, while the camera utilises low-power, low-resolution monochrome cameras. Furthermore, the capabilities of PixelGen extend beyond traditional photography, enabling visualization of phenomena invisible to conventional cameras, such as sound waves. PixelGen can enable numerous novel applications, and we demonstrate that it enables unique visualization of the surroundings that are then projected on extended reality headsets. We believe, PixelGen goes beyond conventional cameras and opens new avenues for research and photography.

This paper introduces Dual Interior Point Learning (DIPL) and Dual Supergradient Learning (DSL) to learn dual feasible solutions to parametric linear programs with bounded variables, which are pervasive across many industries. DIPL mimics a novel dual interior point algorithm while DSL mimics classical dual supergradient ascent. DIPL and DSL ensure dual feasibility by predicting dual variables associated with the constraints then exploiting the flexibility of the duals of the bound constraints. DIPL and DSL complement existing primal learning methods by providing a certificate of quality. They are shown to produce high-fidelity dual-feasible solutions to large-scale optimal power flow problems providing valid dual bounds under 0.5% optimality gap.

Deep learning for time series forecasting has traditionally operated within a one-model-per-dataset framework, limiting its potential to leverage the game-changing impact of large pre-trained models. The concept of universal forecasting, emerging from pre-training on a vast collection of time series datasets, envisions a single Large Time Series Model capable of addressing diverse downstream forecasting tasks. However, constructing such a model poses unique challenges specific to time series data: i) cross-frequency learning, ii) accommodating an arbitrary number of variates for multivariate time series, and iii) addressing the varying distributional properties inherent in large-scale data. To address these challenges, we present novel enhancements to the conventional time series Transformer architecture, resulting in our proposed Masked Encoder-based Universal Time Series Forecasting Transformer (Moirai). Trained on our newly introduced Large-scale Open Time Series Archive (LOTSA) featuring over 27B observations across nine domains, Moirai achieves competitive or superior performance as a zero-shot forecaster when compared to full-shot models. Code, model weights, and data will be released.

In contrast to classical cognitive science which studied brains in isolation, ecological approaches focused on the role of the body and environment in shaping cognition. Similarly, in this thesis we adopt an ecological approach to grounded natural language understanding (NLU) research. Grounded language understanding studies language understanding systems situated in the context of events, actions and precepts in naturalistic/simulated virtual environments. Where classic research tends to focus on designing new models and optimization methods while treating environments as given, we explore the potential of environment design for improving data collection and model development. We developed novel training and annotation approaches for procedural text understanding based on text-based game environments. We also drew upon embodied cognitive linguistics literature to propose a roadmap for grounded NLP research, and to inform the development of a new benchmark for measuring the progress of large language models on challenging commonsense reasoning tasks. We leveraged the richer supervision provided by text-based game environments to develop Breakpoint Transformers, a novel approach to modeling intermediate semantic information in long narrative or procedural texts. Finally, we integrated theories on the role of environments in collective human intelligence to propose a design for AI-augmented "social thinking environments" for knowledge workers like scientists.

Speeding up the large-scale distributed training is challenging in that it requires improving various components of training including load balancing, communication, optimizers, etc. We present novel approaches for fast large-scale training of BERT model which individually ameliorates each component thereby leading to a new level of BERT training performance. Load balancing is imperative in distributed BERT training since its training datasets are characterized by samples with various lengths. Communication cost, which is proportional to the scale of distributed training, needs to be hidden by useful computation. In addition, the optimizers, e.g., ADAM, LAMB, etc., need to be carefully re-evaluated in the context of large-scale distributed training. We propose two new ideas, (1) local presorting based on dataset stratification for load balancing and (2) bucket-wise gradient clipping before allreduce which allows us to benefit from the overlap of gradient computation and synchronization as well as the fast training of gradient clipping before allreduce. We also re-evaluate existing optimizers via hyperparameter optimization and utilize ADAM, which also contributes to fast training via larger batches than existing methods. Our proposed methods, all combined, give the fastest MLPerf BERT training of 25.1 (22.3) seconds on 1,024 NVIDIA A100 GPUs, which is 1.33x (1.13x) and 1.57x faster than the other top two (one) submissions to MLPerf v1.1 (v2.0). Our implementation and evaluation results are available at MLPerf v1.1~v2.1.