Domain generalization (DG) seeks predictors which perform well on unseen test distributions by leveraging data drawn from multiple related training distributions or domains. To achieve this, DG is commonly formulated as an average- or worst-case problem over the set of possible domains. However, predictors that perform well on average lack robustness while predictors that perform well in the worst case tend to be overly-conservative. To address this, we propose a new probabilistic framework for DG where the goal is to learn predictors that perform well with high probability. Our key idea is that distribution shifts seen during training should inform us of probable shifts at test time, which we realize by explicitly relating training and test domains as draws from the same underlying meta-distribution. To achieve probable DG, we propose a new optimization problem called Quantile Risk Minimization (QRM). By minimizing the $\alpha$-quantile of predictor's risk distribution over domains, QRM seeks predictors that perform well with probability $\alpha$. To solve QRM in practice, we propose the Empirical QRM (EQRM) algorithm and provide: (i) a generalization bound for EQRM; and (ii) the conditions under which EQRM recovers the causal predictor as $\alpha \to 1$. In our experiments, we introduce a more holistic quantile-focused evaluation protocol for DG and demonstrate that EQRM outperforms state-of-the-art baselines on datasets from WILDS and DomainBed.

Jared Genser in many ways fits a certain Washington, D.C., type. He wears navy suits and keeps his hair cut short. He graduated from a top law school, joined a large firm, and made partner at 40. Eventually, he became disenchanted with big law and started his own boutique practice with offices off--where else--Dupont Circle. What distinguishes Genser from the city's other 50-something lawyers is his unusual clientele: He represents high-value political prisoners.

Methane is one of the most potent greenhouse gases, and its short atmospheric half-life makes it a prime target to rapidly curb global warming. However, current methane emission monitoring techniques primarily rely on approximate emission factors or self-reporting, which have been shown to often dramatically underestimate emissions. Although initially designed to monitor surface properties, satellite multispectral data has recently emerged as a powerful method to analyze atmospheric content. However, the spectral resolution of multispectral instruments is poor, and methane measurements are typically very noisy. Methane data products are also sensitive to absorption by the surface and other atmospheric gases (water vapor in particular) and therefore provide noisy maps of potential methane plumes, that typically require extensive human analysis. Here, we show that the image recognition capabilities of deep learning methods can be leveraged to automatize the detection of methane leaks in Sentinel-2 satellite multispectral data, with dramatically reduced false positive rates compared with state-of-the-art multispectral methane data products, and without the need for a priori knowledge of potential leak sites. Our proposed approach paves the way for the automated, high-definition and high-frequency monitoring of point-source methane emissions across the world.

Deep neural networks often learn unintended bias during training, which might have harmful effects when deployed in real-world settings. This work surveys 214 papers related to sociodemographic bias in natural language processing (NLP). In this study, we aim to provide a more comprehensive understanding of the similarities and differences among approaches to sociodemographic bias in NLP. To better understand the distinction between bias and real-world harm, we turn to ideas from psychology and behavioral economics to propose a definition for sociodemographic bias. We identify three main categories of NLP bias research: types of bias, quantifying bias, and debiasing techniques. We highlight the current trends in quantifying bias and debiasing techniques, offering insights into their strengths and weaknesses. We conclude that current approaches on quantifying bias face reliability issues, that many of the bias metrics do not relate to real-world bias, and that debiasing techniques need to focus more on training methods. Finally, we provide recommendations for future work.

The primary goal of any company is to increase its profits by improving both the quality of its products and how they are advertised. In this context, neuromarketing seeks to enhance the promotion of products and generate a greater acceptance on potential buyers. Traditionally, neuromarketing studies have relied on a single biosignal to obtain feedback from presented stimuli. However, thanks to new devices and technological advances studying this area of knowledge, recent trends indicate a shift towards the fusion of diverse biosignals. An example is the usage of electroencephalography for understanding the impact of an advertisement at the neural level and visual tracking to identify the stimuli that induce such impacts. This emerging pattern determines which biosignals to employ for achieving specific neuromarketing objectives. Furthermore, the fusion of data from multiple sources demands advanced processing methodologies. Despite these complexities, there is a lack of literature that adequately collates and organizes the various data sources and the applied processing techniques for the research objectives pursued. To address these challenges, the current paper conducts a comprehensive analysis of the objectives, biosignals, and data processing techniques employed in neuromarketing research. This study provides both the technical definition and a graphical distribution of the elements under revision. Additionally, it presents a categorization based on research objectives and provides an overview of the combinatory methodologies employed. After this, the paper examines primary public datasets designed for neuromarketing research together with others whose main purpose is not neuromarketing, but can be used for this matter. Ultimately, this work provides a historical perspective on the evolution of techniques across various phases over recent years and enumerates key lessons learned.

Large language models (LLMs) have shown impressive ability for open-domain NLP tasks. However, LLMs are sometimes too footloose for natural language understanding (NLU) tasks which always have restricted output and input format. Their performances on NLU tasks are highly related to prompts or demonstrations and are shown to be poor at performing several representative NLU tasks, such as event extraction and entity typing. To this end, we present SeqGPT, a bilingual (i.e., English and Chinese) open-source autoregressive model specially enhanced for open-domain natural language understanding. We express all NLU tasks with two atomic tasks, which define fixed instructions to restrict the input and output format but still ``open'' for arbitrarily varied label sets. The model is first instruction-tuned with extremely fine-grained labeled data synthesized by ChatGPT and then further fine-tuned by 233 different atomic tasks from 152 datasets across various domains. The experimental results show that SeqGPT has decent classification and extraction ability, and is capable of performing language understanding tasks on unseen domains. We also conduct empirical studies on the scaling of data and model size as well as on the transfer across tasks. Our model is accessible at https://github.com/Alibaba-NLP/SeqGPT.

Precise and fast prediction methods for ischemic areas comprised of dead tissue, core, and salvageable tissue, penumbra, in acute ischemic stroke (AIS) patients are of significant clinical interest. They play an essential role in improving diagnosis and treatment planning. Computed Tomography (CT) scan is one of the primary modalities for early assessment in patients with suspected AIS. CT Perfusion (CTP) is often used as a primary assessment to determine stroke location, severity, and volume of ischemic lesions. Current automatic segmentation methods for CTP mostly use already processed 3D parametric maps conventionally used for clinical interpretation by radiologists as input. Alternatively, the raw CTP data is used on a slice-by-slice basis as 2D+time input, where the spatial information over the volume is ignored. In addition, these methods are only interested in segmenting core regions, while predicting penumbra can be essential for treatment planning. This paper investigates different methods to utilize the entire 4D CTP as input to fully exploit the spatio-temporal information, leading us to propose a novel 4D convolution layer. Our comprehensive experiments on a local dataset of 152 patients divided into three groups show that our proposed models generate more precise results than other methods explored. Adopting the proposed 4D mJ-Net, a Dice Coefficient of 0.53 and 0.23 is achieved for segmenting penumbra and core areas, respectively. The code is available on https://github.com/Biomedical-Data-Analysis-Laboratory/4D-mJ-Net.git.

Model checking strategic abilities was successfully developed and applied since the early 2000s to ensure properties in Multi-Agent System. In this paper, we introduce the notion of capacities giving different abilities to an agent. This applies naturally to systems where multiple entities can play the same role in the game, such as different client versions in protocol analysis, different robots in heterogeneous fleets, different personality traits in social structure modeling, or different attacker profiles in a cybersecurity setting. With the capacity of other agents being unknown at the beginning of the game, the longstanding problems of imperfect information arise. Our contribution is the following: (i) we define a new class of concurrent game structures where the agents have different capacities that modify their action list and (ii) we introduce a logic extending Alternating-time Temporal Logic to reason about these games.

Deep learning techniques depend on large datasets whose annotation is time-consuming. To reduce annotation burden, the self-training (ST) and active-learning (AL) methods have been developed as well as methods that combine them in an iterative fashion. However, it remains unclear when each method is the most useful, and when it is advantageous to combine them. In this paper, we propose a new method that combines ST with AL using Test-Time Augmentations (TTA). First, TTA is performed on an initial teacher network. Then, cases for annotation are selected based on the lowest estimated Dice score. Cases with high estimated scores are used as soft pseudo-labels for ST. The selected annotated cases are trained with existing annotated cases and ST cases with border slices annotations. We demonstrate the method on MRI fetal body and placenta segmentation tasks with different data variability characteristics. Our results indicate that ST is highly effective for both tasks, boosting performance for in-distribution (ID) and out-of-distribution (OOD) data. However, while self-training improved the performance of single-sequence fetal body segmentation when combined with AL, it slightly deteriorated performance of multi-sequence placenta segmentation on ID data. AL was helpful for the high variability placenta data, but did not improve upon random selection for the single-sequence body data. For fetal body segmentation sequence transfer, combining AL with ST following ST iteration yielded a Dice of 0.961 with only 6 original scans and 2 new sequence scans. Results using only 15 high-variability placenta cases were similar to those using 50 cases. Code is available at: https://github.com/Bella31/TTA-quality-estimation-ST-AL

The 2019-2020 Australia bushfire season was catastrophic in the scale of damage caused to agricultural resources, property, infrastructure, and ecological systems. By the end of 2020, the devastation attributable to these Black Summer fires totalled 33 lives lost, almost 19 million hectares of land burned, over 3,000 homes destroyed and AUD $1.7 billion in insurance losses, as well as an estimated 1 billion animals killed, including half of Kangaroo Island's population of koalas (Filkov et al. 2020). According to the Australian Government Bureau of Meteorology (2021), 2019 was the warmest year on record in Australia, and the period from 2013-2020 represents eight of the ten warmest years in recorded history. There is concern and expectation that impacts of climate change - including more extreme temperatures, persistent drought, and changes in plant growth and landscape drying - will worsen conditions for extreme bushfires (CSIRO and Australian Government Bureau of Meteorology 2020; Deb et al. 2020). Contributing to the problem is that dry lightning represents the main source of natural ignition, and fires that start in remote areas deep in the temperate forests are difficult to access and monitor (Abram et al. 2021). Therefore, opportunities to detect fire ignitions, monitor bushfire spread, and understand movement patterns in remote areas are important for developing effective strategies to mitigate bushfire impact.