When the Chinese firm DeepSeek dropped a large language model called R1 two weeks ago, it sent shock waves through the US tech industry. Not only did R1 match the best of the homegrown competition, it was built for a fraction of the cost--and given away for free. DeepSeek has now suddenly become the company to beat. What exactly did it do to rattle the tech world so fully? And what can we learn from the buzz about what's coming next?

OpenAI claims the tool represents a significant step toward its overarching goal of developing artificial general intelligence (AGI) that matches (or surpasses) human performance. It says that what takes the tool "tens of minutes" would take a human many hours. In response to a single query, such as "Draw me up a competitive analysis between streaming platforms," Deep Research will search the web, analyze the information it encounters, and compile a detailed report that cites its sources. It's also able to draw from files uploaded by users. OpenAI developed Deep Research using the same "chain of thought" reinforcement-learning methods it used to create its o1 multistep reasoning model. But while o1 was designed to focus primarily on mathematics, coding, or other STEM-based tasks, Deep Research can tackle a far broader range of subjects.

Last week, shortly before the start of the Chinese New Year, when much of China shuts down for seven days, the state media saluted DeepSeek, a tech startup whose release of a new low-cost, high-performance artificial-intelligence model, known as R1, prompted a big sell-off in tech stocks on Wall Street. China Central Television showed footage of DeepSeek's bespectacled founder, Liang Wenfeng, meeting with Premier Li Qiang, the second-highest-ranking official in the Chinese government. A few days earlier, China Daily, an English-language news site run by the Chinese Communist Party, had hailed DeepSeek's success, which defied U.S. restrictions on the export of high-performance semiconductor chips used to train A.I. models, as "not an isolated phenomenon, but rather a reflection of the broader vibrancy of China's AI ecosystem." As if to reinforce the point, on Wednesday, the first day of the Year of the Snake, Alibaba, the Chinese tech giant, released its own new A.I. model, which the company claimed "outperforms" competing products from U.S. companies like OpenAI and Meta "almost across the board." Alibaba's claims haven't been independently verified yet, but the DeepSeek-inspired stock sell-off provoked a great deal of commentary about how the company achieved its breakthrough, the durability of U.S. leadership in A.I., and the wisdom of trying to slow down China's tech industry by restricting high-tech exports--a policy that both the first Trump Administration and the Biden Administration followed.

SoftBank Group will spend 3 billion a year to adopt and deploy OpenAI technology throughout its operations, while the two companies have agreed to form a joint venture to market the artificial intelligence as an enterprise solution. "This initiative will not only transform the way SoftBank Group operates but also revolutionize the way companies work in Japan and around the globe," SoftBank CEO Masayoshi Son said in a statement Monday. The technology, which the company describes as an advanced enterprise AI called Cristal intelligence, will be used at all companies under the SoftBank group, including Arm, Line and PayPay, to improve productivity and drive innovation. For instance, SoftBank's telecom unit plans to make more than 100 million workflows automated, the company said in the press release.

Two days after releasing o3-mini to the world, the company made a surprise announcement on Sunday evening, revealing Deep Research. The new feature allows ChatGPT to find, analyze and synthesize hundreds of websites and online sources to create reports "at the level of a research analyst." The chatbot will then take "anywhere from 5 to 30 minutes" to compile an answer, a side panel documenting the agent's progress and citations as it works. "It accomplishes in tens of minutes what would take a human many hours," OpenAI says of the new feature. "Our ultimate aspiration is a model that can uncover and discover new knowledge for itself," said Mark Chen, chief research officer at OpenAI, during the company's reveal livestream.

This tutorial intends to introduce readers with a background in AI to quantum machine learning (QML) -- a rapidly evolving field that seeks to leverage the power of quantum computers to reshape the landscape of machine learning. For self-consistency, this tutorial covers foundational principles, representative QML algorithms, their potential applications, and critical aspects such as trainability, generalization, and computational complexity. In addition, practical code demonstrations are provided in https://qml-tutorial.github.io/ to illustrate real-world implementations and facilitate hands-on learning. Together, these elements offer readers a comprehensive overview of the latest advancements in QML. By bridging the gap between classical machine learning and quantum computing, this tutorial serves as a valuable resource for those looking to engage with QML and explore the forefront of AI in the quantum era.

Information design (ID) explores how a sender influence the optimal behavior of receivers to achieve specific objectives. While ID originates from everyday human communication, existing game-theoretic and machine learning methods often model information structures as numbers, which limits many applications to toy games. This work leverages LLMs and proposes a verbalized framework in Bayesian persuasion (BP), which extends classic BP to real-world games involving human dialogues for the first time. Specifically, we map the BP to a verbalized mediator-augmented extensive-form game, where LLMs instantiate the sender and receiver. To efficiently solve the verbalized game, we propose a generalized equilibrium-finding algorithm combining LLM and game solver. The algorithm is reinforced with techniques including verbalized commitment assumptions, verbalized obedience constraints, and information obfuscation. Numerical experiments in dialogue scenarios, such as recommendation letters, courtroom interactions, and law enforcement, validate that our framework can both reproduce theoretical results in classic BP and discover effective persuasion strategies in more complex natural language and multi-stage scenarios.

We propose a novel family of language models, Latent-Thought Language Models (LTMs), which incorporate explicit latent thought vectors that follow an explicit prior model in latent space. These latent thought vectors guide the autoregressive generation of ground tokens through a Transformer decoder. Training employs a dual-rate optimization process within the classical variational Bayes framework: fast learning of local variational parameters for the posterior distribution of latent vectors, and slow learning of global decoder parameters. Empirical studies reveal that LTMs possess additional scaling dimensions beyond traditional LLMs, yielding a structured design space. Higher sample efficiency can be achieved by increasing training compute per token, with further gains possible by trading model size for more inference steps. Designed based on these scaling properties, LTMs demonstrate superior sample and parameter efficiency compared to conventional autoregressive models and discrete diffusion models. They significantly outperform these counterparts in validation perplexity and zero-shot language modeling. Additionally, LTMs exhibit emergent few-shot in-context reasoning capabilities that scale with model and latent size, and achieve competitive performance in conditional and unconditional text generation.

Large language models (LLMs) have been adopted to solve sequential decision-making tasks such as multi-armed bandits (MAB), in which an LLM is directly instructed to select the arms to pull in every iteration. However, this paradigm of direct arm selection using LLMs has been shown to be suboptimal in many MAB tasks. Therefore, we propose an alternative approach which combines the strengths of classical MAB and LLMs. Specifically, we adopt a classical MAB algorithm as the high-level framework and leverage the strong in-context learning capability of LLMs to perform the sub-task of reward prediction. Firstly, we incorporate the LLM-based reward predictor into the classical Thompson sampling (TS) algorithm and adopt a decaying schedule for the LLM temperature to ensure a transition from exploration to exploitation. Next, we incorporate the LLM-based reward predictor (with a temperature of 0) into a regression oracle-based MAB algorithm equipped with an explicit exploration mechanism. We also extend our TS-based algorithm to dueling bandits where only the preference feedback between pairs of arms is available, which requires non-trivial algorithmic modifications. We conduct empirical evaluations using both synthetic MAB tasks and experiments designed using real-world text datasets, in which the results show that our algorithms consistently outperform previous baseline methods based on direct arm selection. Interestingly, we also demonstrate that in challenging tasks where the arms lack semantic meanings that can be exploited by the LLM, our approach achieves considerably better performance than LLM-based direct arm selection.

We propose SCONE ($\textbf{S}$calable, $\textbf{C}$ontextualized, $\textbf{O}$ffloaded, $\textbf{N}$-gram $\textbf{E}$mbedding), a method for extending input embedding layers to enhance language model performance as layer size scales. To avoid increased decoding costs, SCONE retains the original vocabulary while introducing embeddings for a set of frequent $n$-grams. These embeddings provide contextualized representation for each input token and are learned with a separate model during training. During inference, they are precomputed and stored in off-accelerator memory with minimal impact on inference speed. SCONE enables two new scaling strategies: increasing the number of cached $n$-gram embeddings and scaling the model used to learn them, all while maintaining fixed inference-time FLOPS. We show that scaling both aspects allows SCONE to outperform a 1.9B parameter baseline across diverse corpora, while using only half the inference-time FLOPS.