We present LaMDA: Language Models for Dialog Applications. LaMDA is a family of Transformer-based neural language models specialized for dialog, which have up to 137B parameters and are pre-trained on 1.56T words of public dialog data and web text. While model scaling alone can improve quality, it shows less improvements on safety and factual grounding. We demonstrate that fine-tuning with annotated data and enabling the model to consult external knowledge sources can lead to significant improvements towards the two key challenges of safety and factual grounding. The first challenge, safety, involves ensuring that the model's responses are consistent with a set of human values, such as preventing harmful suggestions and unfair bias. We quantify safety using a metric based on an illustrative set of human values, and we find that filtering candidate responses using a LaMDA classifier fine-tuned with a small amount of crowdworker-annotated data offers a promising approach to improving model safety. The second challenge, factual grounding, involves enabling the model to consult external knowledge sources, such as an information retrieval system, a language translator, and a calculator. We quantify factuality using a groundedness metric, and we find that our approach enables the model to generate responses grounded in known sources, rather than responses that merely sound plausible. Finally, we explore the use of LaMDA in the domains of education and content recommendations, and analyze their helpfulness and role consistency.

Online selection of optimal waveforms for target tracking with active sensors has long been a problem of interest. Many conventional solutions utilize an estimation-theoretic interpretation, in which a waveform-specific Cram\'{e}r-Rao lower bound on measurement error is used to select the optimal waveform for each tracking step. However, this approach is only valid in the high SNR regime, and requires a rather restrictive set of assumptions regarding the target motion and measurement models. Further, due to computational concerns, many traditional approaches are limited to near-term, or myopic, optimization, even though radar scenes exhibit strong temporal correlation. More recently, reinforcement learning has been proposed for waveform selection, in which the problem is framed as a Markov decision process (MDP), allowing for long-term planning. However, a major limitation of reinforcement learning is that the memory length of the underlying Markov process is often unknown for realistic target and channel dynamics, and a more general framework is desirable. This work develops a universal sequential waveform selection scheme which asymptotically achieves Bellman optimality in any radar scene which can be modeled as a $U^{\text{th}}$ order Markov process for a finite, but unknown, integer $U$. Our approach is based on well-established tools from the field of universal source coding, where a stationary source is parsed into variable length phrases in order to build a context-tree, which is used as a probabalistic model for the scene's behavior. We show that an algorithm based on a multi-alphabet version of the Context-Tree Weighting (CTW) method can be used to optimally solve a broad class of waveform-agile tracking problems while making minimal assumptions about the environment's behavior.

With the development of new sensors and monitoring devices, more sources of data become available to be used as inputs for machine learning models. These can on the one hand help to improve the accuracy of a model. On the other hand however, combining these new inputs with historical data remains a challenge that has not yet been studied in enough detail. In this work, we propose a transfer-learning algorithm that combines the new and the historical data, that is especially beneficial when the new data is scarce. We focus the approach on the linear regression case, which allows us to conduct a rigorous theoretical study on the benefits of the approach. We show that our approach is robust against negative transfer-learning, and we confirm this result empirically with real and simulated data.

Parents in Fairfax, Va., protest two controversial books being displayed in the holiday reading section of school libraries. Fairfax County Public Schools (FCPS) is seeking a social media monitoring service that will track hate speech as well as purported harassment and threats against employees, students, or racial groups – resurfacing concerns about First Amendment rights and school safety. A request for proposal (RFP), which closed last week, showed the Virginia school district offering $200,000 to "detect help deter any negative actions or consequences coming from social media which may be directed to racial groups or any student or teacher within FCPS." Under technical and functional requirements, the school district lists "Automatically classify aliases, usernames, emails websites, etc."; "Visually identify relationships and connections between persons"; "Save search queries and set alerts for active listening"; and "False positive reduction with embedded violence language classifier and metadata optimization technology." LINCOLNIA, VIRGINIA - AUGUST 23: Masked students arrive to the first day of class at Glasgow Middle School in Lincolnia, Virginia, on Monday, August 23, 2021, the first day back to school for the Fairfax County school district.

Four Democratic senators and House representatives have called on several government departments to stop using Clearview AI's facial recognition system. The Government Accountability Office said in August that the Departments of Justice, Defense, Homeland Security and the Interior were all using the contentious technology for "domestic law enforcement." Pramila Jayapal and Ayanna Pressley urged the agencies to refrain from using Clearview's products and other facial recognition tools. "Clearview AI's technology could eliminate public anonymity in the United States," the lawmakers wrote to the agencies in their letters, which were obtained by The Verge. They said that, combined with the facial recognition system, the database of billions of photos Clearview scraped from social media platforms "is capable of fundamentally dismantling Americans' expectation that they can move, assemble or simply appear in public without being identified."

We've seen again and again the harmful, unintended consequences of irresponsibly deployed algorithms: risk assessment tools in the criminal justice system amplifying racial discrimination, false arrests powered by facial recognition, massive environmental costs of server farms, unacknowledged psychological harm from social media interactions, and new, sometimes-insurmountable hurdles in accessing public services. These actual harms are egregious, but what makes the current regime hopeless is that companies are incentivized to remain ignorant (or at least claim they to be) about the harms they expose us to, lest they be found liable. Many of the current ideas for regulating large tech companies won't address this ignorance or the harms it causes. While proposed antitrust laws would reckon with harms emerging from diminished competition in the digital markets, relatively small companies can also have disturbing, far-reaching power to affect our lives. Even if these proposed regulatory tools were to push tech companies away from some harmful practices, researchers, advocates and--critically --communities affected by these practices would still not have sufficient say in all the ways these companies' algorithms shape our lives.

The Senate Judiciary committee holds a hearing on Wednesday on U.S. military drone strikes. The event is scheduled to begin at 10 a.m. This is a developing story and will be updated.

Artificial Intelligence (AI) is becoming more prevalent in all of our lives. The so-called "intelligence" is the result of powerful computers sorting through mountains of data to find patterns, using algorithms designed and optimized by computer scientists. Like all technology, AI is far from perfect. As we have started using AI for consequential decisions, we have realized that while AI can improve decision making, it too often compounds historical patterns of bias and deepens existing inequality. AI's reliance on biased data or design processes has led to systems that produce discriminatory, or otherwise harmful, outcomes.

Biases in artificial intelligence will aim to be eradicated in a world first as the NHS in England trials a new approach to the ethical adoption of AI in healthcare. Algorithmic Impact Assessment (AIA), designed by the Ada Lovelace Institute, will be piloted to support researchers and developers to assess the possible risks and biases of AI systems to patients and the public before they can access NHS data. While artificial intelligence has the potential to support health and care workers to deliver better care for people, it could also exacerbate existing health inequalities if concerns such as algorithmic bias aren't accounted for. While AI has great potential to transform health and care services, we must tackle biases which have the potential to do further harm to some populations as part of our mission to eradicate health disparities. This pilot once again demonstrates the UK is at the forefront of adopting new technologies in a way that is ethical and patient-centred.

The EU strikes again with a new set of regulations that take aim at the use of artificial intelligence (AI) to address the variety of risks associated with the societal adoption of AI. Like its sibling the General Data Protection Regulation (GDPR), the Artificial Intelligence Act (AIA) actually has teeth, with fines rising to €30 million, or 6% of global revenue. Is the answer to delete all your AI systems to minimize your risk to zero, or continue using AI for a competitive edge? Can you manage the recurring costs required to maintain compliance with the AIA even as the technology itself increases your bottomline? Take the famous UK pub chain JD Wetherspoon, founded by British businessman Tim Martin in 1979 who has been an outspoken critic of the EU and a Brexit campaigner. Their response to personal identifiable information (PII) protection, legislated by the GDPR in 2017, was to delete their entire CRM database.