Improving the reasoning capabilities of large language models (LLMs) has attracted considerable interest. Recent approaches primarily focus on improving the reasoning process to yield a more precise final answer. However, in scenarios involving contextually aware reasoning, these methods neglect the importance of first identifying logical relationships from the context before proceeding with the reasoning. This oversight could lead to a superficial understanding and interaction with the context, potentially undermining the quality and reliability of the reasoning outcomes. In this paper, we propose an information re-organization (InfoRE) method before proceeding with the reasoning to enhance the reasoning ability of LLMs. Our re-organization method involves initially extracting logical relationships from the contextual content, such as documents or paragraphs, and subsequently pruning redundant content to minimize noise. Then, we utilize the re-organized information in the reasoning process. This enables LLMs to deeply understand the contextual content by clearly perceiving these logical relationships, while also ensuring high-quality responses by eliminating potential noise. To demonstrate the effectiveness of our approach in improving the reasoning ability, we conduct experiments using Llama2-70B, GPT-3.5, and GPT-4 on various contextually aware multi-hop reasoning tasks. Using only a zero-shot setting, our method achieves an average absolute improvement of 4% across all tasks, highlighting its potential to improve the reasoning performance of LLMs. Our source code is available at https://github.com/hustcxx/InfoRE.

In this paper, we introduce PoseCrafter, a one-shot method for personalized video generation following the control of flexible poses. Built upon Stable Diffusion and ControlNet, we carefully design an inference process to produce high-quality videos without the corresponding ground-truth frames. First, we select an appropriate reference frame from the training video and invert it to initialize all latent variables for generation. Then, we insert the corresponding training pose into the target pose sequences to enhance faithfulness through a trained temporal attention module. Furthermore, to alleviate the face and hand degradation resulting from discrepancies between poses of training videos and inference poses, we implement simple latent editing through an affine transformation matrix involving facial and hand landmarks. Extensive experiments on several datasets demonstrate that PoseCrafter achieves superior results to baselines pre-trained on a vast collection of videos under 8 commonly used metrics. Besides, PoseCrafter can follow poses from different individuals or artificial edits and simultaneously retain the human identity in an open-domain training video.

Combining semantic information with behavioral data is a crucial research area in recommender systems. A promising approach involves leveraging external knowledge to enrich behavioral-based recommender systems with abundant semantic information. However, this approach faces two primary challenges: denoising raw external knowledge and adapting semantic representations. To address these challenges, we propose an External Knowledge-Enhanced Recommendation method with LLM Assistance (TRAWL). This method utilizes large language models (LLMs) to extract relevant recommendation knowledge from raw external data and employs a contrastive learning strategy for adapter training. Experiments on public datasets and real-world online recommender systems validate the effectiveness of our approach.

Controlled Text Generation (CTG) aims to produce texts that exhibit specific desired attributes. In this study, we introduce a pluggable CTG framework for Large Language Models (LLMs) named Dynamic Attribute Graphs-based controlled text generation (DATG). This framework utilizes an attribute scorer to evaluate the attributes of sentences generated by LLMs and constructs dynamic attribute graphs. DATG modulates the occurrence of key attribute words and key anti-attribute words, achieving effective attribute control without compromising the original capabilities of the model. We conduct experiments across four datasets in two tasks: toxicity mitigation and sentiment transformation, employing five LLMs as foundational models. Our findings highlight a remarkable enhancement in control accuracy, achieving a peak improvement of 19.29% over baseline methods in the most favorable task across four datasets. Additionally, we observe a significant decrease in perplexity, markedly improving text fluency.

Moreover, the program committee comprised researchers from 12 countries across five continents. The workshop featured six keynote speakers, oral sessions, poster sessions, a panel discussion, and a networking lunch. Of the 20 submitted papers, six were selected for oral and poster presentation, and an additional nine were selected for poster presentation only. The acceptance rate was, therefore, 75%. All accepted papers are published in the open-access workshop's proceedings at https://ceur-ws.org/Vol-3649/.

Meta has seen strikingly little AI-generated misinformation around the 2024 elections despite major votes in countries such as Indonesia, Taiwan, and Bangladesh, said the company's president of global affairs, Nick Clegg, on Wednesday. "The interesting thing so far--I stress, so far--is not how much but how little AI-generated content [there is]," said Clegg during an interview at MIT Technology Review's EmTech Digital conference in Cambridge, Massachusetts. As voters will head to polls this year in more than 50 countries, experts have raised the alarm over AI-generated political disinformation and the prospect that malicious actors will use generative AI and social media to interfere with elections. And even well-resourced tech giants like Meta are struggling to keep up. How generative AI is boosting the spread of disinformation and propaganda.

This article was featured in the One Story to Read Today newsletter. Let's get this out of the way: OpenAI's voice assistant doesn't sound that much like Scarlett Johansson. The movie star has alleged that, though she rebuffed multiple attempts by Sam Altman, the company's CEO, to license her voice for the product that it demoed last week, the one it ended up using was "eerily similar" to her own. Not everyone finds the similarity so eerie--to my ear, it lacks her distinctive smoky rasp--but at the very least, the new AI does appear to imitate the playful lilts and cadences that Johansson used while playing Samantha, the digital assistant in the 2013 film Her. That's depressing--and not only because OpenAI may have run roughshod over Johansson's wishes, but because it has made such an unimaginative choice.

The Pixel 8a starts at 499 ( 549/ 499/A 849). That may be 50 more than last year's 7a, but the new model improves just about everything, and undercuts the Pixel 8 by 200. Google has revamped the design of the phone, giving it a more rounded shape, particularly at the corners, which makes it nicer to hold without a case. It is still unmistakably a Pixel, with a big aluminium camera bar across the back, and looks great in one of its more bold colours. The 6.1in screen has been upgraded from last year to match the more expensive Pixel 8, and is significantly brighter, running at up to 120Hz for smooth scrolling. It is a crisp and colourful display at a good size.

OpenAI cast the actor of Sky's voice months before Sam Altman contacted Scarlett Johansson, and it had no intention of finding someone who sounded like her, according to The Washington Post. The publication said the flier OpenAI issued last year looked for actors that had "warm, engaging [and] charismatic" voices. They needed to be between 25 and 45 years old and had to be non-union, but OpenAI reportedly didn't specify that it was looking for a Scarlett Johansson voice-alike. If you'll recall, Johansson accused the company of copying her likeness without permission for its Sky voice assistant. The agent of Sky's voice told The Post that the company never talked about Johansson or the movie Her with their talent.

This article addresses the societal costs associated with the lack of regulation in Artificial Intelligence and proposes a framework combining innovation and regulation. Over fifty years of AI research, catalyzed by declining computing costs and the proliferation of data, have propelled AI into the mainstream, promising significant economic benefits. Yet, this rapid adoption underscores risks, from bias amplification and labor disruptions to existential threats posed by autonomous systems. The discourse is polarized between "accelerationists," advocating for unfettered technological advancement, and "doomers," calling for a slowdown to prevent dystopian outcomes. This piece advocates for a middle path that leverages technical innovation and smart regulation to maximize AI's potential benefits while minimizing its risks, offering a pragmatic approach to the responsible progress of AI technology. Technical invention beyond today's most capable foundation models is needed to contain catastrophic risks. Regulation is required to create incentives for this research while addressing current issues.