When I was growing up, I watched a lot of sci-fi movies about aliens that come to Earth. The extraterrestrials in popular culture, however, always looked so familiar that I found them far-fetched. What are the chances that E.T., the Predator, or ALF would develop arms and legs, a humanlike face, and opposable thumbs? Perhaps as a result, I associated alien life more with fantasy than with science, and I never gave much thought to what a visit would really look like. But my attitude started to change in 2020, when I read Liu Cixin's "The Three-Body Problem" and its two sequels.

To mitigate the potential misuse of large language models (LLMs), recent research has developed watermarking algorithms, which restrict the generation process to leave an invisible trace for watermark detection. Due to the two-stage nature of the task, most studies evaluate the generation and detection separately, thereby presenting a challenge in unbiased, thorough, and applicable evaluations. In this paper, we introduce WaterBench, the first comprehensive benchmark for LLM watermarks, in which we design three crucial factors: (1) For \textbf{benchmarking procedure}, to ensure an apples-to-apples comparison, we first adjust each watermarking method's hyper-parameter to reach the same watermarking strength, then jointly evaluate their generation and detection performance. (2) For \textbf{task selection}, we diversify the input and output length to form a five-category taxonomy, covering $9$ tasks. (3) For \textbf{evaluation metric}, we adopt the GPT4-Judge for automatically evaluating the decline of instruction-following abilities after watermarking. We evaluate $4$ open-source watermarks on $2$ LLMs under $2$ watermarking strengths and observe the common struggles for current methods on maintaining the generation quality. The code and data are available at \url{https://github.com/THU-KEG/WaterBench}.

To design effective vaccine policies, policymakers need detailed data about who has been vaccinated, who is holding out, and why. However, existing data in the US are insufficient: reported vaccination rates are often delayed or missing, and surveys of vaccine hesitancy are limited by high-level questions and self-report biases. Here, we show how large-scale search engine logs and machine learning can be leveraged to fill these gaps and provide novel insights about vaccine intentions and behaviors. First, we develop a vaccine intent classifier that can accurately detect when a user is seeking the COVID-19 vaccine on search. Our classifier demonstrates strong agreement with CDC vaccination rates, with correlations above 0.86, and estimates vaccine intent rates to the level of ZIP codes in real time, allowing us to pinpoint more granular trends in vaccine seeking across regions, demographics, and time. To investigate vaccine hesitancy, we use our classifier to identify two groups, vaccine early adopters and vaccine holdouts. We find that holdouts, compared to early adopters matched on covariates, are 69% more likely to click on untrusted news sites. Furthermore, we organize 25,000 vaccine-related URLs into a hierarchical ontology of vaccine concerns, and we find that holdouts are far more concerned about vaccine requirements, vaccine development and approval, and vaccine myths, and even within holdouts, concerns vary significantly across demographic groups. Finally, we explore the temporal dynamics of vaccine concerns and vaccine seeking, and find that key indicators emerge when individuals convert from holding out to preparing to accept the vaccine.

What use could healthcare have for someone who makes things up, can't keep a secret, doesn't really know anything, and, when speaking, simply fills in the next word based on what's come before? Lots, if that individual is the newest form of artificial intelligence, according to some of the biggest companies out there. Companies pushing the latest AI technology -- known as "generative AI" -- are piling on: Google and Microsoft want to bring types of so-called large language models to healthcare. Big firms that are familiar to folks in white coats -- but maybe less so to your average Joe and Jane -- are equally enthusiastic: Electronic medical records giants Epic and Oracle Cerner aren't far behind. The space is crowded with startups, too.

Large language models (LLMs) transfer well to new tasks out-of-the-box simply given a natural language prompt that demonstrates how to perform the task and no additional training. Prompting is a brittle process wherein small modifications to the prompt can cause large variations in the model predictions, and therefore significant effort is dedicated towards designing a painstakingly "perfect prompt" for a task. To mitigate the high degree of effort involved in prompt-design, we instead ask whether producing multiple effective, yet imperfect, prompts and aggregating them can lead to a high quality prompting strategy. Our observations motivate our proposed prompting method, ASK ME ANYTHING (AMA). We first develop an understanding of the effective prompt formats, finding that question-answering (QA) prompts, which encourage open-ended generation ("Who went to the park?") tend to outperform those that restrict the model outputs ("John went to the park. Output True or False."). Our approach recursively uses the LLM itself to transform task inputs to the effective QA format. We apply the collected prompts to obtain several noisy votes for the input's true label. We find that the prompts can have very different accuracies and complex dependencies and thus propose to use weak supervision, a procedure for combining the noisy predictions, to produce the final predictions for the inputs. We evaluate AMA across open-source model families (e.g., EleutherAI, BLOOM, OPT, and T0) and model sizes (125M-175B parameters), demonstrating an average performance lift of 10.2% over the few-shot baseline. This simple strategy enables the open-source GPT-J-6B model to match and exceed the performance of few-shot GPT3-175B on 15 of 20 popular benchmarks. Averaged across these tasks, the GPT-J-6B model outperforms few-shot GPT3-175B. We release our code here: https://github.com/HazyResearch/ama_prompting

The companies racing to deploy autonomous cars on the world's roads took a reality check in the 2010s, but multimillion-dollar development efforts remain ongoing across the automotive and tech industries. German supplier Bosch is notably moving full speed ahead with its quest to make driverless cars a reality. Kay Stepper, Bosch's senior vice president of automated driving, sat down with Digital Trends to talk about the state of autonomous driving in 2020, and what's next for the artificial intelligence technology that powers the prototypes it's testing. Bosch has never made a car, so it brings its innovations to the market through partnerships with automakers. It chose Mercedes-Benz parent company Daimler to test autonomous technology in real-world conditions via a ridesharing pilot program in San Jose, California, close to one of the company's research centers.

In this thesis we discuss machine learning methods performing automated variable selection for learning sparse predictive models. There are multiple reasons for promoting sparsity in the predictive models. By relying on a limited set of input variables the models naturally counteract the overfitting problem ubiquitous in learning from finite sets of training points. Sparse models are cheaper to use for predictions, they usually require lower computational resources and by relying on smaller sets of inputs can possibly reduce costs for data collection and storage. Sparse models can also contribute to better understanding of the investigated phenomenons as they are easier to interpret than full models.

Artificial intelligence, robots, and automation are rising in importance in many areas. As noted in the recent book, "The Future of Work: Robots, AI, and Automation," there are exciting advances in finance, transportation, national defense, smart cities, and health care, among other areas. Businesses are developing solutions that improve the efficiency and effectiveness of their operations and using these tools to improve the way their firms function. Yet there also are concerns about the impact of these developments on jobs and personal privacy. A Pew Research Center national survey revealed considerable unease about emerging trends.

We present a new method for forecasting systems of multiple interrelated time series. The method learns the forecast models together with discovering leading indicators from within the system that serve as good predictors improving the forecast accuracy and a cluster structure of the predictive tasks around these. The method is based on the classical linear vector autoregressive model (VAR) and links the discovery of the leading indicators to inferring sparse graphs of Granger causality. We formulate a new constrained optimisation problem to promote the desired sparse structures across the models and the sharing of information amongst the learning tasks in a multi-task manner. We propose an algorithm for solving the problem and document on a battery of synthetic and real-data experiments the advantages of our new method over baseline VAR models as well as the state-of-the-art sparse VAR learning methods.