Language models (LMs) have been instrumental for the rapid advance of natural language processing. This paper studies continual pre-training of LMs, in particular, continual domain-adaptive pre-training (or continual DAP-training). Existing research has shown that further pre-training an LM using a domain corpus to adapt the LM to the domain can improve the end-task performance in the domain. This paper proposes a novel method to continually DAP-train an LM with a sequence of unlabeled domain corpora to adapt the LM to these domains to improve their end-task performances. The key novelty of our method is a soft-masking mechanism that directly controls the update to the LM. A novel proxy is also proposed to preserve the general knowledge in the original LM. Additionally, it contrasts the representations of the previously learned domain knowledge (including the general knowledge in the pre-trained LM) and the knowledge from the current full network to achieve knowledge integration. The method not only overcomes catastrophic forgetting, but also achieves knowledge transfer to improve end-task performances. Empirical evaluation demonstrates the effectiveness of the proposed method.

A method of near real-time detection and tracking of resident space objects (RSOs) using a convolutional neural network (CNN) and linear quadratic estimator (LQE) is proposed. Advances in machine learning architecture allow the use of low-power/cost embedded devices to perform complex classification tasks. In order to reduce the costs of tracking systems, a low-cost embedded device will be used to run a CNN detection model for RSOs in unresolved images captured by a gray-scale camera and small telescope. Detection results computed in near real-time are then passed to an LQE to compute tracking updates for the telescope mount, resulting in a fully autonomous method of optical RSO detection and tracking. Keywords: Space Domain Awareness, Neural Networks, Real-Time, Object Detection, Embedded Systems.

Audiovisual representation learning typically relies on the correspondence between sight and sound. However, there are often multiple audio tracks that can correspond with a visual scene. Consider, for example, different conversations on the same crowded street. The effect of such counterfactual pairs on audiovisual representation learning has not been previously explored. To investigate this, we use dubbed versions of movies to augment cross-modal contrastive learning. Our approach learns to represent alternate audio tracks, differing only in speech content, similarly to the same video. Our results show that dub-augmented training improves performance on a range of auditory and audiovisual tasks, without significantly affecting linguistic task performance overall. We additionally compare this approach to a strong baseline where we remove speech before pretraining, and find that dub-augmented training is more effective, including for paralinguistic and audiovisual tasks where speech removal leads to worse performance. These findings highlight the importance of considering speech variation when learning scene-level audiovisual correspondences and suggest that dubbed audio can be a useful augmentation technique for training audiovisual models toward more robust performance.

Note-level automatic music transcription is one of the most representative music information retrieval (MIR) tasks and has been studied for various instruments to understand music. However, due to the lack of high-quality labeled data, transcription of many instruments is still a challenging task. In particular, in the case of singing, it is difficult to find accurate notes due to its expressiveness in pitch, timbre, and dynamics. In this paper, we propose a method of finding note onsets of singing voice more accurately by leveraging the linguistic characteristics of singing, which are not seen in other instruments. The proposed model uses mel-scaled spectrogram and phonetic posteriorgram (PPG), a frame-wise likelihood of phoneme, as an input of the onset detection network while PPG is generated by the pre-trained network with singing and speech data. To verify how linguistic features affect onset detection, we compare the evaluation results through the dataset with different languages and divide onset types for detailed analysis. Our approach substantially improves the performance of singing transcription and therefore emphasizes the importance of linguistic features in singing analysis.

Natural language processing (NLP) has a significant impact on society via technologies such as machine translation and search engines. Despite its success, NLP technology is only widely available for high-resource languages such as English and Chinese, while it remains inaccessible to many languages due to the unavailability of data resources and benchmarks. In this work, we focus on developing resources for languages in Indonesia. Despite being the second most linguistically diverse country, most languages in Indonesia are categorized as endangered and some are even extinct. We develop the first-ever parallel resource for 10 low-resource languages in Indonesia. Our resource includes datasets, a multi-task benchmark, and lexicons, as well as a parallel Indonesian-English dataset. We provide extensive analyses and describe the challenges when creating such resources. We hope that our work can spark NLP research on Indonesian and other underrepresented languages.

Sign language recognition (SLR) is a weakly supervised task that annotates sign videos as textual glosses. Recent studies show that insufficient training caused by the lack of large-scale available sign datasets becomes the main bottleneck for SLR. Most SLR works thereby adopt pretrained visual modules and develop two mainstream solutions. The multi-stream architectures extend multi-cue visual features, yielding the current SOTA performances but requiring complex designs and might introduce potential noise. Alternatively, the advanced single-cue SLR frameworks using explicit cross-modal alignment between visual and textual modalities are simple and effective, potentially competitive with the multi-cue framework. In this work, we propose a novel contrastive visual-textual transformation for SLR, CVT-SLR, to fully explore the pretrained knowledge of both the visual and language modalities. Based on the single-cue cross-modal alignment framework, we propose a variational autoencoder (VAE) for pretrained contextual knowledge while introducing the complete pretrained language module. The VAE implicitly aligns visual and textual modalities while benefiting from pretrained contextual knowledge as the traditional contextual module. Meanwhile, a contrastive cross-modal alignment algorithm is designed to explicitly enhance the consistency constraints. Extensive experiments on public datasets (PHOENIX-2014 and PHOENIX-2014T) demonstrate that our proposed CVT-SLR consistently outperforms existing single-cue methods and even outperforms SOTA multi-cue methods.

Modeling and control of epidemics such as the novel Corona virus have assumed paramount importance at a global level. A natural and powerful dynamical modeling framework to use in this context is a continuous time Markov decision process (CTMDP) that encompasses classical compartmental paradigms such as the Susceptible-Infected-Recovered (SIR) model. The challenges with CTMDP based models motivate the need for a more efficient approach and the mean field approach offers an effective alternative. The mean field approach computes the collective behavior of a dynamical system comprising numerous interacting nodes (where nodes represent individuals in the population). This paper (a) presents an overview of the mean field approach to epidemic modeling and control and (b) provides a state-of-the-art update on recent advances on this topic. Our discussion in this paper proceeds along two specific threads. The first thread assumes that the individual nodes faithfully follow a socially optimal control policy prescribed by a regulatory authority. The second thread allows the individual nodes to exhibit independent, strategic behavior. In this case, the strategic interaction is modeled as a mean field game and the control is based on the associated mean field Nash equilibria. In this paper, we start with a discussion of modeling of epidemics using an extended compartmental model - SIVR and provide an illustrative example. We next provide a review of relevant literature, using a mean field approach, on optimal control of epidemics, dealing with how a regulatory authority may optimally contain epidemic spread in a population. Following this, we provide an update on the literature on the use of the mean field game based approach in the study of epidemic spread and control. We conclude the paper with relevant future research directions.

A lot of what Conor Grennan does as a dean of students at NYU's Stern School of Business could be done at least in part by bots. Brainstorming and planning are prime examples of tasks that can be easily handled by generative AI tools like ChatGPT. But instead of feeling like he could be replaced by AI, Grennan has become an evangelist of this technology and its potential to make work better. He likens the opportunity to work with AI technology right now to finding material wealth. "It feels like the Gold Rush, like there's a bunch of people getting to California and seeing little flakes of gold in the river," he told Vox.

Much of the conversation around artificial intelligence (AI) these days centers on whether it will eventually take your job, how it's trying to compete with humans in creative fields, or how it can be misused, say, as a writing tool. You can probably chalk this one-sidedness up to an all-too-human tendency to be suspicious of new tech that isn't well understood by the mainstream (yet). But AI isn't intrinsically evil or good: It's a tool, a vast technology with enormous potential, and there are myriad ways to implement it beyond the current discourse. One vitally important use case is helping us fight and survive the consequences of climate change. Whether it's mitigating the effects of disasters such as floods and fires more quickly or building a cleaner energy grid, the evidence is mounting that AI has an essential role to play in helping to protect us as the planet reacts to climate change. And we'll need all the help we can get.

Abstract--Graph mining algorithms have been playing a significant role in myriad fields over the years. However, despite their promising performance on various graph analytical tasks, most of these algorithms lack fairness considerations. As a consequence, they could lead to discrimination towards certain populations when exploited in human-centered applications. Recently, algorithmic fairness has been extensively studied in graph-based applications. In contrast to algorithmic fairness on independent and identically distributed (i.i.d.) data, fairness in graph mining has exclusive backgrounds, taxonomies, and fulfilling techniques. In this survey, we provide a comprehensive and up-to-date introduction of existing literature under the context of fair graph mining. Specifically, we propose a novel taxonomy of fairness notions on graphs, which sheds light on their connections and differences. We further present an organized summary of existing techniques that promote fairness in graph mining. Finally, we discuss current research challenges and open questions, aiming at encouraging cross-breeding ideas and further advances. Graph-structured data is pervasive in diverse real-world Compared with achieving fairness in the context of independent applications, e.g., E-commerce [102], [121], health care [37], and identically distributed (i.i.d.) data, fulfilling [53], traffic forecasting [72], [100], and drug discovery [15], fairness in graph mining can be non-trivial due to two [172]. The first challenge is to formulate proper have been proposed to gain a deeper understanding of such fairness notions as the criteria to determine the existence of data. These algorithms have shown promising performance unfairness (i.e., bias). Although a vast amount of traditional on graph analytical tasks such as node classification [59], algorithmic fairness notions have been proposed centered [86], [161] and link prediction [4], [103], [109], contributing on i.i.d. For example, the same population can be most of them lack fairness considerations. Consequently, connected with different topologies as in Figure 1a and 1b, they could yield discriminatory results towards certain populations where each node represents an individual, and the color when such algorithms are exploited in humancentered of nodes denotes their demographic subgroup membership, applications [80]. Compared with the graph topology job recommender system may unfavorably recommend in Figure 1a, the topology in Figure 1b has more intra-group fewer job opportunities to individuals of a certain edges than inter-group edges. The dominance of intra-group gender [97] or individuals in an underrepresented ethnic edges in the graph topology is a common type of bias group [150].