These days, it seems like there's nothing AI programs can't do. Thanks to advancements in artificial intelligence, deepfakes have done digital "face-offs" with Hollywood celebrities in films and TV shows, VFX artists can de-age actors almost instantly, and ChatGPT has learned how to write big-budget screenplays in the blink of an eye. Pretty soon, AI will probably decide who wins at the Oscars. Within the past year, AI has also been used to generate beautiful works of art in seconds, creating a viral new trend and causing a boon for fan artists everywhere. TikTok user @cyborgism recently broke the internet by posting a clip featuring many AI-generated pictures of Breaking Bad. The theme here is that the characters are depicted as anime characters straight out of the 1980s, and the result is concerning to say the least.

The recent progress in text-based audio retrieval was largely propelled by the release of suitable datasets. Since the manual creation of such datasets is a laborious task, obtaining data from online resources can be a cheap solution to create large-scale datasets. We study the recently proposed SoundDesc benchmark dataset, which was automatically sourced from the BBC Sound Effects web page. In our analysis, we find that SoundDesc contains several duplicates that cause leakage of training data to the evaluation data. This data leakage ultimately leads to overly optimistic retrieval performance estimates in previous benchmarks. We propose new training, validation, and testing splits for the dataset that we make available online. To avoid weak contamination of the test data, we pool audio files that share similar recording setups. In our experiments, we find that the new splits serve as a more challenging benchmark.

Knowledge-driven dialog system has recently made remarkable breakthroughs. Compared with general dialog systems, superior knowledge-driven dialog systems can generate more informative and knowledgeable responses with pre-provided knowledge. However, in practical applications, the dialog system cannot be provided with corresponding knowledge in advance because it cannot know in advance the development of the conversation. Therefore, in order to make the knowledge dialogue system more practical, it is vital to find a way to retrieve relevant knowledge based on the dialogue history. To solve this problem, we design a knowledge-driven dialog system named DRKQG (Dynamically Retrieving Knowledge via Query Generation for informative dialog response). Specifically, the system can be divided into two modules: the query generation module and the dialog generation module. First, a time-aware mechanism is utilized to capture context information, and a query can be generated for retrieving knowledge through search engine. Then, we integrate the copy mechanism and transformers, which allows the response generation module to produce responses derived from the context and retrieved knowledge. Experimental results at LIC2022, Language and Intelligence Technology Competition, show that our module outperforms the baseline model by a large margin on automatic evaluation metrics, while human evaluation by the Baidu Linguistics team shows that our system achieves impressive results in Factually Correct and Knowledgeable.

This paper presents a new approach to the FNC-1 fake news classification task which involves employing pre-trained encoder models from similar NLP tasks, namely sentence similarity and natural language inference, and two neural network architectures using this approach are proposed. Methods in data augmentation are explored as a means of tackling class imbalance in the dataset, employing common pre-existing methods and proposing a method for sample generation in the under-represented class using a novel sentence negation algorithm. Comparable overall performance with existing baselines is achieved, while significantly increasing accuracy on an under-represented but nonetheless important class for FNC-1.

Decision-makers in the humanitarian sector rely on timely and exact information during crisis events. Knowing how many civilians were injured during an earthquake is vital to allocate aids properly. Information about such victim counts is often only available within full-text event descriptions from newspapers and other reports. Extracting numbers from text is challenging: numbers have different formats and may require numeric reasoning. This renders purely string matching-based approaches insufficient. As a consequence, fine-grained counts of injured, displaced, or abused victims beyond fatalities are often not extracted and remain unseen. We cast victim count extraction as a question answering (QA) task with a regression or classification objective. We compare regex, dependency parsing, semantic role labeling-based approaches, and advanced text-to-text models. Beyond model accuracy, we analyze extraction reliability and robustness which are key for this sensitive task. In particular, we discuss model calibration and investigate few-shot and out-of-distribution performance. Ultimately, we make a comprehensive recommendation on which model to select for different desiderata and data domains. Our work is among the first to apply numeracy-focused large language models in a real-world use case with a positive impact.

With the accelerated technology adoption by a growing number of users, social media have become the main medium for the dissemination of information on current news and events. While these new media bring several benefits (e.g., a large number of consumers reached, instant and continuous updates on one's topics of interest), they also enable the spread of harmful information in the form of fake news, and may thus polarize public discourse regarding critical topics (e.g., elections [32], vaccination [30], health hazards [24]) and threaten democratic values [35]. Because of its detrimental effects on society at large, the fake news phenomenon has been studied by scientists and practitioners alike; fake news is defined as news articles that intentionally contain verifiably false misleading information inconsistent with factual reality [2, 23, 46, 10, 4, 13, 43]. To mitigate the threat of fake news, journalists have started to manually classify news and offer websites with fact-checking mechanisms that provide a verdict regarding its veracity, such as PolitiFact (https://www.politifact.com/)

ChatGPT, the cutting-edge chatbot from OpenAI that was released in November 2022, can solve math equations, write a history term paper, compose a sonnet and almost everything in between. So it's not surprising that many educators support banning the chatbot in schools to prevent plagiarism, cheating and just plain inaccuracy. In response to these concerns, some major districts have banned the chatbot in schools. In December, the Los Angeles Unified School District "preemptively" blocked access to ChatGPT while "a risk/benefit assessment is conducted," a district spokesperson told the Washington Post. And in January, New York City Public Schools banned access to ChatGPT from devices and networks that the school owns, per the Washington Post.

Anthropic, an AI startup co-founded by former employees of OpenAI, has quietly begun testing a new, ChatGPT-like AI assistant named Claude. The team at Anthropic was gracious enough to grant us access, and updates to Anthropic's social media policies mean we can now share some of our early, informal comparison findings between Claude and ChatGPT. To show how Claude is different, we'll begin by asking ChatGPT and Claude to introduce themselves with the same prompt. Short and to the point -- ChatGPT is an assistant made to answer questions and sound human. The interface to Claude is a Slack channel using a bot that edits messages to make text appear word-by-word. This causes "(edited)" to appear.

For the last five or so years, Qualcomm has bet big on bringing more artificial intelligence to smartphones, laptops, vehicles, smart infrastructure and other devices in the field--or what the company calls the "connected intelligent edge." It's Don McGuire's job to tell Qualcomm's technology and artificial intelligence story in a way that's not scary. Recently, that's been harder to do. Last fall's launch of ChatGPT--a generative AI chatbot that answers prompts with polished essays, poetry, computer code and other human-like content--has thrust artificial intelligence into the public spotlight, with decidedly mixed reactions. While there's been plenty of positive hype, many people view the launch of ChatGPT--and AI overall --with a good amount of hand-wringing.

Get Simon Calder's Travel email Fast, intelligent, cheap: ChatGPT – the AI chatbot system capable of spewing out facts like a caffeinated Stephen Fry – is the hot new thing on the block that's here to claim everything you hold dear. Or so it seems to a slew of journalists who have begun questioning their credentials now big tech is here to do what they do best – except faster and for less money. In recent months, we've seen the loquacious creation firing out answers to life's big questions, writing haikus, job applications and even producing a university paper in 20 minutes and bagging a 2:2 grade in the process. With its seemingly infinite ability to regurgitate facts about everything from Jan Morris to Mauritian cuisine, some journalists have begun to worry that their jobs might be at risk. Lisa Gibbs, the director of news partnerships at the Associated Press, noted in a December Google News Initiative talk that while "robots are not the journalists of the future – they are a journalist's assistant, a very good one", she added that her organisation could "find news faster and break news faster" with the aid of AI. Elsewhere, Reuters has used an in-house AI programme called Lynx Insight since 2018 and The Washington Post has produced machine-written snippets of copy using its in-house robot report, Heliograf.