We introduce a new open information extraction (OIE) benchmark for pre-trained language models (LM). Recent studies have demonstrated that pre-trained LMs, such as BERT and GPT, may store linguistic and relational knowledge. In particular, LMs are able to answer ``fill-in-the-blank'' questions when given a pre-defined relation category. Instead of focusing on pre-defined relations, we create an OIE benchmark aiming to fully examine the open relational information present in the pre-trained LMs. We accomplish this by turning pre-trained LMs into zero-shot OIE systems. Surprisingly, pre-trained LMs are able to obtain competitive performance on both standard OIE datasets (CaRB and Re-OIE2016) and two new large-scale factual OIE datasets (TAC KBP-OIE and Wikidata-OIE) that we establish via distant supervision. For instance, the zero-shot pre-trained LMs outperform the F1 score of the state-of-the-art supervised OIE methods on our factual OIE datasets without needing to use any training sets. Our code and datasets are available at https://github.com/cgraywang/IELM

In hate speech detection, developing training and evaluation datasets across various domains is the critical issue. Whereas, major approaches crawl social media texts and hire crowd-workers to annotate the data. Following this convention often restricts the scope of pejorative expressions to a single domain lacking generalization. Sometimes domain overlap between training corpus and evaluation set overestimate the prediction performance when pretraining language models on low-data language. To alleviate these problems in Korean, we propose APEACH that asks unspecified users to generate hate speech examples followed by minimal post-labeling. We find that APEACH can collect useful datasets that are less sensitive to the lexical overlaps between the pretraining corpus and the evaluation set, thereby properly measuring the model performance.

In a fair world, people have equitable opportunities to education, to conduct scientific research, to publish, and to get credit for their work, regardless of where they live. However, it is common knowledge among researchers that a vast number of papers accepted at top NLP venues come from a handful of western countries and (lately) China; whereas, very few papers from Africa and South America get published. Similar disparities are also believed to exist for paper citation counts. In the spirit of "what we do not measure, we cannot improve", this work asks a series of questions on the relationship between geographical location and publication success (acceptance in top NLP venues and citation impact). We first created a dataset of 70,000 papers from the ACL Anthology, extracted their meta-information, and generated their citation network. We then show that not only are there substantial geographical disparities in paper acceptance and citation but also that these disparities persist even when controlling for a number of variables such as venue of publication and sub-field of NLP. Further, despite some steps taken by the NLP community to improve geographical diversity, we show that the disparity in publication metrics across locations is still on an increasing trend since the early 2000s. We release our code and dataset here: https://github.com/iamjanvijay/acl-cite-net

This work demonstrates that Legal Judgement Prediction systems without expert-informed adjustments can be vulnerable to shallow, distracting surface signals that arise from corpus construction, case distribution, and confounding factors. To mitigate this, we use domain expertise to strategically identify statistically predictive but legally irrelevant information. We adopt adversarial training to prevent the system from relying on it. We evaluate our deconfounded models by employing interpretability techniques and comparing to expert annotations. Quantitative experiments and qualitative analysis show that our deconfounded model consistently aligns better with expert rationales than baselines trained for prediction only. We further contribute a set of reference expert annotations to the validation and testing partitions of an existing benchmark dataset of European Court of Human Rights cases.

The recent literature in text classification is biased towards short text sequences (e.g., sentences or paragraphs). In real-world applications, multi-page multi-paragraph documents are common and they cannot be efficiently encoded by vanilla Transformer-based models. We compare different Transformer-based Long Document Classification (TrLDC) approaches that aim to mitigate the computational overhead of vanilla transformers to encode much longer text, namely sparse attention and hierarchical encoding methods. We examine several aspects of sparse attention (e.g., size of local attention window, use of global attention) and hierarchical (e.g., document splitting strategy) transformers on four document classification datasets covering different domains. We observe a clear benefit from being able to process longer text, and, based on our results, we derive practical advice of applying Transformer-based models on long document classification tasks.

We often hear of the dark side of artificial intelligence (AI), how it will plunge us into a dystopian world of lost privacy and bad automated decisions, culminating in the robots killing us all. Professor Orly Lobel's The Equality Machine: Harnessing Digital Technology for a Brighter, More Inclusive Future (Public Affairs, October 2022) offers a very different view – one of optimism. Orly's book is an exuberant and insightful account of the bright side of AI and related digital technologies. Her book is filled with fascinating facts and engaging stories. Orly Lobel is the Warren Distinguished Professor of Law; University Professor; and Director, Center for Employment and Labor Policy at the U.C. San Diego School of Law.

Artificial intelligence laws and regulations are still a new concept for most countries, and the current state of rules is neither universal nor inclusive. It's crystal clear that AI's constructive and negative roles in every sector require a call for agreed-upon rules. Artificial intelligence (AI), the creation of computer systems that can learn and make decisions without the need for human intelligence, has the potential to revolutionize and foster innovation in business and government. More goods and services are entering the market as AI research and technology continue to advance. For instance, businesses are creating AI to help people manage their houses and let the elderly live in their homes for longer. AI is applied in many aspects of modern life, including self-driving cars, digital assistants, and healthcare technology. However, initiatives to look into and create standards have been motivated by worries about potential abuse or unforeseen repercussions of AI. Already, AI is enhancing healthcare, connecting people in new ways, and drastically increasing productivity. But when applied incorrectly or irresponsibly, AI might result in employment losses, prejudiced or racist outcomes, etc.

In the discussion paper, the UK financial supervisory authorities have not provided a new legal framework or their intended future approaches for regulating the use of AI and machine learning in financial services. However, they have assessed the benefits, risks and harms related to the use of AI, and the current legal framework that applies to AI in financial services. The UK financial services regulators, the Bank of England (BoE), the Prudential Regulation Authority (PRA) and the Financial Conduct Authority (FCA) (together Supervisory Authorities) jointly published a discussion paper (DP5/22) on artificial intelligence (AI) and machine learning on 11 October 2022. The purpose of the discussion paper was to facilitate a public debate on the safe and responsible adoption of AI in UK financial services. The Supervisory Authorities have also raised discussion questions for stakeholder input, with the aim of understanding whether the current regulatory framework is sufficient to address the potential risks and harms associated with AI and how any additional intervention may support the safe and responsible adoption of AI in UK financial services.

The role of artificial intelligence (AI) in regulation is changing rapidly, as the technology matures and is increasingly applied across a broad range of industries. While AI has the potential to improve regulatory outcomes by complementing human decision-making, there are also risks associated with its use. As AI technologies become more ubiquitous, it is important for regulators to understand both the opportunities and challenges posed by AI. AI can be used for a variety of tasks in regulation, from automating repetitive tasks to providing recommendations on how to respond to complex situations. Automation can free up resources that can be redirected to other activities, such as risk management or policy analysis. Moreover, by analyzing large data sets, AI can identify patterns that humans may not be able to detect.

Humor is an inherently social phenomenon, with humorous utterances shaped by what is socially and culturally accepted. Understanding humor is an important NLP challenge, with many applications to human-computer interactions. In this work we explore humor in the context of Cards Against Humanity -- a party game where players complete fill-in-the-blank statements using cards that can be offensive or politically incorrect. We introduce a novel dataset of 300,000 online games of Cards Against Humanity, including 785K unique jokes, analyze it and provide insights. We trained machine learning models to predict the winning joke per game, achieving performance twice as good (20\%) as random, even without any user information. On the more difficult task of judging novel cards, we see the models' ability to generalize is moderate. Interestingly, we find that our models are primarily focused on punchline card, with the context having little impact. Analyzing feature importance, we observe that short, crude, juvenile punchlines tend to win.