The San Francisco police could use robots to deploy lethal force under a policy advanced by city supervisors on Tuesday that thrust the city into the forefront of a national debate about the use of weaponized robots in American cities. The possibility is not merely hypothetical. In 2016, the Dallas Police Department ended a standoff with a gunman suspected of killing five officers by blowing him up with a bomb attached to a robot in what was believed to be the first lethal use of the technology by an American law enforcement agency. Supporters of the policy, advanced by the San Francisco Board of Supervisors by an 8-to-3 vote, said it would allow the police to deploy a robot with deadly force in extraordinary circumstances, such as when a mass shooter or a terrorist is threatening the lives of officers or civilians. David Lazar, assistant chief of the San Francisco Police Department, cited as an example the gunman who opened fire from his Las Vegas high-rise hotel room in 2017, killing 60 people in the deadliest mass shooting in modern American history.

A week ago, the San Francisco Police Department (SFPD) petitioned the Board of Supervisors for permission to deploy robots that can kill suspects under specific circumstances. Now, the board has approved the petition with a vote of 8 vs. 3 despite strong opposition from civil liberties groups. Under the new policy, robots can be used "as a deadly force option when risk of loss of life to members of the public or officers are imminent and outweigh any other force option available to the SFPD." The city's police force has over a dozen robots at the moment, which are equipped with the capability to provide video reconnaissance and to diffuse bombs. None of them have weapons and live ammunition, the SFPD says, and there are no plans to fit them with any.

Independent algorithmic auditing firm Parity AI has partnered with talent acquisition and management platform Beamery to conduct ongoing scrutiny of bias in its artificial intelligence (AI) hiring tools. Beamery, which uses AI to help businesses identify, recruit, develop, retain and redeploy talent, approached Parity to conduct a third-party audit of its systems, which was completed in early November 2022. To accompany the audit, Beamery has also published an accompanying "explainability statement" outlining its commitment to responsible AI. Liz O'Sullivan, CEO of Parity, says there is a "significant challenge" for businesses and human resources (HR) teams in reassuring all stakeholders involved that their AI tools are privacy-conscious and do not discriminate against disadvantaged or marginalised communities. "To do this, businesses must be able to demonstrate that their systems comply with all relevant regulations, including local, federal and international human rights, civil rights and data protection laws," she says. "We are delighted to work with the Beamery team as an example of a company that genuinely cares about minimising unintentional algorithmic bias, in order to serve their communities well.

Officials in San Francisco have voted to give the city's police the power to use potentially lethal, remote-controlled robots in emergency situations. The 8-3 vote in favour of the move followed an emotionally charged two-hour debate and came despite strong objections from civil liberties and other police oversight groups in the city on the west coast of the United States. Supervisor Connie Chan, a member of the committee that forwarded the proposal to the full board, said she understood concerns over use of force but that "according to state law, we are required to approve the use of these equipments. So here we are, and it's definitely not an easy discussion." The San Francisco Police Department (SFPD) has said it does not have pre-armed robots and has no plans to arm robots with guns.

Police in San Francisco will be allowed to deploy potentially lethal, remote-controlled robots in emergency situations. The controversial policy was approved after weeks of scrutiny and a heated debate among the city's board of supervisors during their meeting on Tuesday. Police oversight groups, the ACLU and San Francisco's public defender had urged the 11-member body to reject the police's use of equipment proposal. Opponents of the policy said it would lead to further militarization of a police force already too aggressive with underserved communities. They said the parameters under which use would be allowed were too vague.

Concept relatedness estimation (CRE) aims to determine whether two given concepts are related. Existing methods only consider the pairwise relationship between concepts, while overlooking the higher-order relationship that could be encoded in a concept-level graph structure. We discover that this underlying graph satisfies a set of intrinsic properties of CRE, including reflexivity, commutativity, and transitivity. In this paper, we formalize the CRE properties and introduce a graph structure named ConcreteGraph. To address the data scarcity issue in CRE, we introduce a novel data augmentation approach to sample new concept pairs from the graph. As it is intractable for data augmentation to fully capture the structural information of the ConcreteGraph due to a large amount of potential concept pairs, we further introduce a novel Graph Component Contrastive Learning framework to implicitly learn the complete structure of the ConcreteGraph. Empirical results on three datasets show significant improvement over the state-of-the-art model. Detailed ablation studies demonstrate that our proposed approach can effectively capture the high-order relationship among concepts.

Question Answering (QA) is a growing area of research, often used to facilitate the extraction of information from within documents. State-of-the-art QA models are usually pre-trained on domain-general corpora like Wikipedia and thus tend to struggle on out-of-domain documents without fine-tuning. We demonstrate that synthetic domain-specific datasets can be generated easily using domain-general models, while still providing significant improvements to QA performance. We present two new tools for this task: A flexible pipeline for validating the synthetic QA data and training downstream models on it, and an online interface to facilitate human annotation of this generated data. Using this interface, crowdworkers labelled 1117 synthetic QA pairs, which we then used to fine-tune downstream models and improve domain-specific QA performance by 8.75 F1.

Synthetic data generation has recently gained widespread attention as a more reliable alternative to traditional data anonymization. The involved methods are originally developed for image synthesis. Hence, their application to the typically tabular and relational datasets from healthcare, finance and other industries is non-trivial. While substantial research has been devoted to the generation of realistic tabular datasets, the study of synthetic relational databases is still in its infancy. In this paper, we combine the variational autoencoder framework with graph neural networks to generate realistic synthetic relational databases. We then apply the obtained method to two publicly available databases in computational experiments. The results indicate that real databases' structures are accurately preserved in the resulting synthetic datasets, even for large datasets with advanced data types.

Conventional closed-world information extraction (IE) approaches rely on human ontologies to define the scope for extraction. As a result, such approaches fall short when applied to new domains. This calls for systems that can automatically infer new types from given corpora, a task which we refer to as type discovery. To tackle this problem, we introduce the idea of type abstraction, where the model is prompted to generalize and name the type. Then we use the similarity between inferred names to induce clusters. Observing that this abstraction-based representation is often complementary to the entity/trigger token representation, we set up these two representations as two views and design our model as a co-training framework. Our experiments on multiple relation extraction and event extraction datasets consistently show the advantage of our type abstraction approach. Code available at https://github.com/raspberryice/type-discovery-abs.

Pretrained transformer models have achieved state-of-the-art results in many tasks and benchmarks recently. Many state-of-the-art Language Models (LMs), however, do not scale well above the threshold of 512 input tokens. In specialized domains though (such as legal, scientific or biomedical), models often need to process very long text (sometimes well above 10000 tokens). Even though many efficient transformers have been proposed (such as Longformer, BigBird or FNet), so far, only very few such efficient models are available for specialized domains. Additionally, since the pretraining process is extremely costly in general - but even more so as the sequence length increases - it is often only in reach of large research labs. One way of making pretraining cheaper is the Replaced Token Detection (RTD) task, by providing more signal during training, since the loss can be computed over all tokens. In this work, we train Longformer models with the efficient RTD task on legal data to showcase that pretraining efficient LMs is possible using much less compute. We evaluate the trained models on challenging summarization tasks requiring the model to summarize long texts to show to what extent the models can achieve good performance on downstream tasks. We find that both the small and base models outperform their baselines on the in-domain BillSum and out-of-domain PubMed tasks in their respective parameter range. We publish our code and models for research purposes.