Pre-trained language models achieve superior performance but are computationally expensive. Techniques such as pruning and knowledge distillation have been developed to reduce their sizes and latencies. In this work, we propose a structured pruning method GRAIN (Gradient-based Intra-attention pruning), which performs task-specific pruning with knowledge distillation and yields highly effective models. Different from common approaches that prune each attention head as a whole, GRAIN inspects and prunes intra-attention structures, which greatly expands the structure search space and enables more flexible models. We also propose a gradient separation strategy that reduces the interference of distillation on pruning for a better combination of the two approaches. Experiments on GLUE, SQuAD, and CoNLL 2003 show that GRAIN notably outperforms other methods, especially in the high sparsity regime, and achieves $6\sim7\times$ speedups while maintaining $93\%\sim99\%$ performance. Under extreme compression where only $3\%$ transformer weights remain, the pruned model is still competitive compared to larger models.

Large language models are trained in two stages: (1) unsupervised pretraining from raw text, to learn general-purpose representations, and (2) large scale instruction tuning and reinforcement learning, to better align to end tasks and user preferences. We measure the relative importance of these two stages by training LIMA, a 65B parameter LLaMa language model fine-tuned with the standard supervised loss on only 1,000 carefully curated prompts and responses, without any reinforcement learning or human preference modeling. LIMA demonstrates remarkably strong performance, learning to follow specific response formats from only a handful of examples in the training data, including complex queries that range from planning trip itineraries to speculating about alternate history. Moreover, the model tends to generalize well to unseen tasks that did not appear in the training data. In a controlled human study, responses from LIMA are either equivalent or strictly preferred to GPT-4 in 43% of cases; this statistic is as high as 58% when compared to Bard and 65% versus DaVinci003, which was trained with human feedback. Taken together, these results strongly suggest that almost all knowledge in large language models is learned during pretraining, and only limited instruction tuning data is necessary to teach models to produce high quality output.

Tensor completion is important to many areas such as computer vision, data analysis, and signal processing. Enforcing low-rank structures on completed tensors, a category of methods known as low-rank tensor completion, has recently been studied extensively. Whilst such methods attained great success, none considered exploiting numerical priors of tensor elements. Ignoring numerical priors causes loss of important information regarding the data, and therefore prevents the algorithms from reaching optimal accuracy. This work attempts to construct a new methodological framework called GCDTC (Generalized CP Decomposition Tensor Completion) for leveraging numerical priors and achieving higher accuracy in tensor completion. In this newly introduced framework, a generalized form of CP Decomposition is applied to low-rank tensor completion. This paper also proposes an algorithm known as SPTC (Smooth Poisson Tensor Completion) for nonnegative integer tensor completion as an instantiation of the GCDTC framework. A series of experiments on real-world data indicate that SPTC could produce results superior in completion accuracy to current state-of-the-art methods. Related code is available in the supplemental materials.

The rise of cryptocurrencies like Bitcoin, which enable transactions with a degree of pseudonymity, has led to a surge in various illicit activities, including ransomware payments and transactions on darknet markets. These illegal activities often utilize Bitcoin as the preferred payment method. However, current tools for detecting illicit behavior either rely on a few heuristics and laborious data collection processes or employ computationally inefficient graph neural network (GNN) models that are challenging to interpret. To overcome the computational and interpretability limitations of existing techniques, we introduce an effective solution called Chainlet Orbits. This approach embeds Bitcoin addresses by leveraging their topological characteristics in transactions. By employing our innovative address embedding, we investigate e-crime in Bitcoin networks by focusing on distinctive substructures that arise from illicit behavior. The results of our node classification experiments demonstrate superior performance compared to state-of-the-art methods, including both topological and GNN-based approaches. Moreover, our approach enables the use of interpretable and explainable machine learning models in as little as 15 minutes for most days on the Bitcoin transaction network.

People in Los Angeles and Austin, Texas, told Fox News whether they think its important for the government to regulate artificial intelligence technology. Fox News Digital dominated April 2023 to finish as the No. 1 news organization in key measurables. Americans flocked to Fox News Digital for the latest news and analysis during the news-heavy month to help it top all competitors in both multiplatform minutes and multiplatform views, according to Comscore. It marked the seventh-straight month as the top news brand among multiplatform views and the 26th straight month as No.1 with multiplatform minutes. April featured exclusive political reporting such as the White House dismissing claims the Biden administration mishandled the federal investigation into Hunter Biden, ex-Director of National Intelligence John Ratcliffe saying the "only explanation" for the COVID-19 global pandemic is a lab leak and former President Donald Trump suing his ex-lawyer Michael Cohen for more than $500 million.

Deep neural networks have consistently shown great performance in several real-world use cases like autonomous vehicles, satellite imaging, etc., effectively leveraging large corpora of labeled training data. However, learning unbiased models depends on building a dataset that is representative of a diverse range of realistic scenarios for a given task. This is challenging in many settings where data comes from high-volume streams, with each scenario occurring in random interleaved episodes at varying frequencies. We study realistic streaming settings where data instances arrive in and are sampled from an episodic multi-distributional data stream. Using submodular information measures, we propose STREAMLINE, a novel streaming active learning framework that mitigates scenario-driven slice imbalance in the working labeled data via a three-step procedure of slice identification, slice-aware budgeting, and data selection. We extensively evaluate STREAMLINE on real-world streaming scenarios for image classification and object detection tasks. We observe that STREAMLINE improves the performance on infrequent yet critical slices of the data over current baselines by up to $5\%$ in terms of accuracy on our image classification tasks and by up to $8\%$ in terms of mAP on our object detection tasks.

We introduce HATELEXICON, a lexicon of slurs and targets of hate speech for the countries of Brazil, Germany, India and Kenya, to aid training and interpretability of models. We demonstrate how our lexicon can be used to interpret model predictions, showing that models developed to classify extreme speech rely heavily on target words when making predictions. Further, we propose a method to aid shot selection for training in low-resource settings via HATELEXICON. In few-shot learning, the selection of shots is of paramount importance to model performance. In our work, we simulate a few-shot setting for German and Hindi, using HASOC data for training and the Multilingual HateCheck (MHC) as a benchmark. We show that selecting shots based on our lexicon leads to models performing better on MHC than models trained on shots sampled randomly. Thus, when given only a few training examples, using our lexicon to select shots containing more sociocultural information leads to better few-shot performance.

Reconstruction error-based neural architectures constitute a classical deep learning approach to anomaly detection which has shown great performances. It consists in training an Autoencoder to reconstruct a set of examples deemed to represent the normality and then to point out as anomalies those data that show a sufficiently large reconstruction error. Unfortunately, these architectures often become able to well reconstruct also the anomalies in the data. This phenomenon is more evident when there are anomalies in the training set. In particular when these anomalies are labeled, a setting called semi-supervised, the best way to train Autoencoders is to ignore anomalies and minimize the reconstruction error on normal data. The goal of this work is to investigate approaches to allow reconstruction error-based architectures to instruct the model to put known anomalies outside of the domain description of the normal data. Specifically, our strategy exploits a limited number of anomalous examples to increase the contrast between the reconstruction error associated with normal examples and those associated with both known and unknown anomalies, thus enhancing anomaly detection performances. The experiments show that this new procedure achieves better performances than the standard Autoencoder approach and the main deep learning techniques for semi-supervised anomaly detection.

In this multicultural age, language translation is one of the most performed tasks, and it is becoming increasingly AI-moderated and automated. As a novel AI system, ChatGPT claims to be proficient in such translation tasks and in this paper, we put that claim to the test. Specifically, we examine ChatGPT's accuracy in translating between English and languages that exclusively use gender-neutral pronouns. We center this study around Bengali, the 7$^{th}$ most spoken language globally, but also generalize our findings across five other languages: Farsi, Malay, Tagalog, Thai, and Turkish. We find that ChatGPT perpetuates gender defaults and stereotypes assigned to certain occupations (e.g. man = doctor, woman = nurse) or actions (e.g. woman = cook, man = go to work), as it converts gender-neutral pronouns in languages to `he' or `she'. We also observe ChatGPT completely failing to translate the English gender-neutral pronoun `they' into equivalent gender-neutral pronouns in other languages, as it produces translations that are incoherent and incorrect. While it does respect and provide appropriately gender-marked versions of Bengali words when prompted with gender information in English, ChatGPT appears to confer a higher respect to men than to women in the same occupation. We conclude that ChatGPT exhibits the same gender biases which have been demonstrated for tools like Google Translate or MS Translator, as we provide recommendations for a human centered approach for future designers of AIs that perform language translation to better accommodate such low-resource languages.

In this paper, we present ISI-Clear, a state-of-the-art, cross-lingual, zero-shot event extraction system and accompanying user interface for event visualization & search. Using only English training data, ISI-Clear makes global events available on-demand, processing user-supplied text in 100 languages ranging from Afrikaans to Yiddish. We provide multiple event-centric views of extracted events, including both a graphical representation and a document-level summary. We also integrate existing cross-lingual search algorithms with event extraction capabilities to provide cross-lingual event-centric search, allowing English-speaking users to search over events automatically extracted from a corpus of non-English documents, using either English natural language queries (e.g. cholera outbreaks in Iran) or structured queries (e.g. find all events of type Disease-Outbreak with agent cholera and location Iran).