A large body of NLP research has documented the ways gender biases manifest and amplify within large language models (LLMs), though this research has predominantly operated within a gender binary-centric context. A growing body of work has identified the harmful limitations of this gender-exclusive framing; many LLMs cannot correctly and consistently refer to persons outside the gender binary, especially if they use neopronouns. While data scarcity has been identified as a possible culprit, the precise mechanisms through which it influences LLM misgendering remain underexplored. Our work addresses this gap by studying data scarcity's role in subword tokenization and, consequently, the formation of LLM word representations. We uncover how the Byte-Pair Encoding (BPE) tokenizer, a backbone for many popular LLMs, contributes to neopronoun misgendering through out-of-vocabulary behavior. We introduce pronoun tokenization parity (PTP), a novel approach to reduce LLM neopronoun misgendering by preserving a token's functional structure. We evaluate PTP's efficacy using pronoun consistency-based metrics and a novel syntax-based metric. Through several controlled experiments, finetuning LLMs with PTP improves neopronoun consistency from 14.5% to 58.4%, highlighting the significant role tokenization plays in LLM pronoun consistency.

Recently, instruction-following audio-language models have received broad attention for audio interaction with humans. However, the absence of pre-trained audio models capable of handling diverse audio types and tasks has hindered progress in this field. Consequently, most existing works have only been able to support a limited range of interaction capabilities. In this paper, we develop the Qwen-Audio model and address this limitation by scaling up audio-language pre-training to cover over 30 tasks and various audio types, such as human speech, natural sounds, music, and songs, to facilitate universal audio understanding abilities. However, directly co-training all tasks and datasets can lead to interference issues, as the textual labels associated with different datasets exhibit considerable variations due to differences in task focus, language, granularity of annotation, and text structure. To overcome the one-to-many interference, we carefully design a multi-task training framework by conditioning on a sequence of hierarchical tags to the decoder for encouraging knowledge sharing and avoiding interference through shared and specified tags respectively. Remarkably, Qwen-Audio achieves impressive performance across diverse benchmark tasks without requiring any task-specific fine-tuning, surpassing its counterparts. Building upon the capabilities of Qwen-Audio, we further develop Qwen-Audio-Chat, which allows for input from various audios and text inputs, enabling multi-turn dialogues and supporting various audio-central scenarios.

The use of NLP in the realm of financial technology is broad and complex, with applications ranging from sentiment analysis and named entity recognition to question answering. Large Language Models (LLMs) have been shown to be effective on a variety of tasks; however, no LLM specialized for the financial domain has been reported in literature. In this work, we present BloombergGPT, a 50 billion parameter language model that is trained on a wide range of financial data. We construct a 363 billion token dataset based on Bloomberg's extensive data sources, perhaps the largest domain-specific dataset yet, augmented with 345 billion tokens from general purpose datasets. We validate BloombergGPT on standard LLM benchmarks, open financial benchmarks, and a suite of internal benchmarks that most accurately reflect our intended usage. Our mixed dataset training leads to a model that outperforms existing models on financial tasks by significant margins without sacrificing performance on general LLM benchmarks. Additionally, we explain our modeling choices, training process, and evaluation methodology. We release Training Chronicles (Appendix C) detailing our experience in training BloombergGPT.

Reliability measures associated with the prediction of the machine learning models are critical to strengthening user confidence in artificial intelligence. Therefore, those models that are able to provide not only predictions, but also reliability, enjoy greater popularity. In the field of recommender systems, reliability is crucial, since users tend to prefer those recommendations that are sure to interest them, that is, high predictions with high reliabilities. In this paper, we propose Restricted Bernoulli Matrix Factorization (ResBeMF), a new algorithm aimed at enhancing the performance of classification-based collaborative filtering. The proposed model has been compared to other existing solutions in the literature in terms of prediction quality (Mean Absolute Error and accuracy scores), prediction quantity (coverage score) and recommendation quality (Mean Average Precision score). The experimental results demonstrate that the proposed model provides a good balance in terms of the quality measures used compared to other recommendation models.

We present VideoPoet, a language model capable of synthesizing high-quality video, with matching audio, from a large variety of conditioning signals. VideoPoet employs a decoder-only transformer architecture that processes multimodal inputs -- including images, videos, text, and audio. The training protocol follows that of Large Language Models (LLMs), consisting of two stages: pretraining and task-specific adaptation. During pretraining, VideoPoet incorporates a mixture of multimodal generative objectives within an autoregressive Transformer framework. The pretrained LLM serves as a foundation that can be adapted for a range of video generation tasks. We present empirical results demonstrating the model's state-of-the-art capabilities in zero-shot video generation, specifically highlighting VideoPoet's ability to generate high-fidelity motions. Project page: http://sites.research.google/videopoet/

Artificial intelligence began to reshape music, movies and art in 2023, sparking both enthusiasm and panic. Some artists used AI to aid their creative practices. Others took legal action against the companies that co-opted art to make their models more powerful. As battles played out across picket lines and courtrooms, millions of viewers and listeners around the world tuned into AI-created content with curiosity, disdain and glee. Here are the major ways AI impacted culture this year.

Fox News contributor Dr. Marc Siegel weighs in on how artificial intelligence can change the patient-doctor relationship on'America's Newsroom.' Artificial intelligence has been able to beat masters at games like chess and poker and Go. AI has also been able to beat human competitors in various video games. While impressive nonetheless, there is one major capability that these games do not require of the AI: physical skill. CyberRunner is an AI tasked with learning how to play the popular labyrinth maze game.

A group of scientists from across the U.S. claim to have created the first artificial intelligence capable of generating AI without human supervision. AI GIVES BIRTH TO AI: Scientists say machine intelligence now capable of replicating without humans. AI (Artificial Intelligence) letters are placed on computer motherboard in this illustration taken on June 23, 2023. TECH AIDING CLIENTS: How AI can brush the dust off the old wealth management industry. AI GONE AWRY: FTC bans Rite Aid's AI facial recognition over lack of consumer protections.

The time has come: our list of the 20 best games of 2023 is now live. I can't remember a year with such an embarrassment of riches to choose from, and the diversity of this list really reflects that. Most outlets – and players – appear to have divided themselves along the lines of Team Baldur's Gate, Team Zelda or Team Alan Wake 2, and any one of them would be a worthy GOTY. In the end you have to go with your heart. Have a read and see if your feelings align with ours.

Toward the end of Baldur's Gate 3, widely considered the most outstanding video game released this year, you can literally go to hell. If you do, you'll have a showdown with the game's equivalent of the devil, a charismatic yet demonic trickster who calls himself Raphael. Naturally, developer Larian Studios wanted it to feel monumental. So they decided that the battle should be accompanied by a song, and that Raphael should be the one singing it. "The idea for a song to be performed by Raphael himself came from our director Swen Vincke about six months before the release of the game," says Borislav Slavov, Baldur's Gate 3's music director.