-- While large language models have rapidly evolved towards general artificial intelligence, their versatility in analyzing time series data remains limited. To address this limitation, we propose a novel normalization technique that considers the inherent nature of tokenization. The proposed Tokenization-Optimized Normalization (TOKON) simplifies time series data by representing each element with a single token, effectively reducing the number of tokens by 2 to 3 times. Additionally, we introduce a novel prompt for time series forecasting, termed Time Series Forecasting with Care (TFSC), to further enhance forecasting performance. Experimental results demonstrate that TOKON improves root mean square error (RMSE) for multi-step forecasting by approximately 7% to 18%, depending on the dataset and prompting method. With the evolution of deep learning in natural language processing, the ubiquitous nature of large language models (LLMs) is becoming increasingly robust [1]. Initially introduced for Q&A services, these models can now be applied to sound and image modalities, enabling LLMs to understand and integrate multi-modal information [2]. The extensive scale of these models allows them to achieve state-of-the-art (SOTA) performance across various tasks, including natural language understanding, event recognition, and coding.

Former House Speaker Kevin McCarthy discusses how the establishment is responding to the Trump admin's shakeup in Washington, D.C. and Transportation Secretary Sean Duffy firing back at'swamp creature' Hillary Clinton. FIRST ON FOX: A top House Republican is moving to make it harder for China to procure advanced U.S. technology amid longstanding concerns about intellectual property theft by Beijing. "My proposed legislation will establish safeguards to prevent future shocks like China's development of DeepSeek using American technology. In addition to the chips China reportedly stockpiled, it appears China used chips under the current export control threshold to achieve this AI breakthrough," House Homeland Security Committee Chairman Mark Green, R-Tenn., told Fox News Digital. "This scenario should be a wakeup call -- if you give the CCP an inch, it will take a mile. The CCP's craftiness is coupled with a total disregard for legal and security considerations. We already know that the CCP uses technology to oppress its own citizens and to commit acts of espionage and sabotage against the United States, including major cyberattacks."

What a great idea I had for the first Plaintext of 2025. After following the frantic competition between OpenAI, Google, Meta, and Anthropic to churn out brainier and deeper "frontier" foundation models, I settled on a thesis about what's ahead: In the new year, those mighty trailblazers will consume billions of dollars, countless gigawatts, and all the silicon Nvidia can muster in their pursuit of AGI. We'll be bombarded by press releases boasting advanced reasoning, more tokens, and maybe even assurances that their models won't make up crazy facts. But people are tired of hearing about how AI is transformational and seeing few transformations to their day-to-day existence. Getting an AI summary of Google search results or having Facebook ask if you want to pose a follow-up question on a post doesn't make you a traveler to the neo-human future.

Summary and Contributions: --- Update after rebuttal --- I thank the author for their detailed rebuttal and effort to clarify the content of the paper and provide missing details. Authors have addressed most pressing concerns, and it is my opinion that their work could be of interest to the community. I would strongly recommend, however, that authors revise the presentation of their manuscript, in particular with respect to clarity/missing details and claims. Please revise/refine the use of certain terms (cf claims about generative models/self-training, see correctness section) and add all the clarifications provided in the rebuttal (in particular with regards to experimental details not provided in the main paper). The method uses the dense attribute attention method of [10] (DAZLE) to learn a set of attribute specific feature vectors, and subsequently train a classification model by iteratively updating classifier (learning from seen and generated unseen feature) and generating new unseen features using classification predictions.

Initially, this paper received diverging reviews. The reviewers found the idea interesting but had some concerns regarding clarity and the difference between the proposed method and the DAZLE baseline. The authors provided a rebuttal, clarifying the issues that were brought up by the reviews, which satisfied the reviewers. During the discussion, some reviewers have argued that the difference between DAZLE and the paper is clear, and the generated features have been demonstrated to have potential to identify new classes. All reviewers have rated the paper as positive (three "6:marginally above threshold" and one "7:accept") after the discussion phase, so overall the reviewers lean toward accepting.

Weaknesses: My main concern is that results seem to be contradictory to what the authors claimed as the benefit of leveraging GPU accelerations. Specifically, in the "impact statement" the authors described CuLE can "provide access to an accelerated training environment to researchers with limited computational capabilities," but the results show the acceleration won't take into effect unless you use more computation---Figure 2, CuLE runs slower than OpenAI when using a fewer number of environments. If someone can only afford to run 100 environments, would this mean CuLE is not useful here? The limitation of the memory has been noted in the paper which is good. I was confused when looking at Table 3. First, why is there no 120 envs experiment for CuLE?

Solving complex planning problems requires Large Language Models (LLMs) to explicitly model the state transition to avoid rule violations, comply with constraints, and ensure optimality-a task hindered by the inherent ambiguity of natural language. To overcome such ambiguity, Planning Domain Definition Language (PDDL) is leveraged as a planning abstraction that enables precise and formal state descriptions. With PDDL, we can generate a symbolic world model where classic searching algorithms, such as A*, can be seamlessly applied to find optimal plans. However, directly generating PDDL domains with current LLMs remains an open challenge due to the lack of PDDL training data. To address this challenge, we propose to scale up the test-time computation of LLMs to enhance their PDDL reasoning capabilities, thereby enabling the generation of high-quality PDDL domains. Specifically, we introduce a simple yet effective algorithm, which first employs a Best-of-N sampling approach to improve the quality of the initial solution and then refines the solution in a fine-grained manner with verbalized machine learning. Our method outperforms o1-mini by a considerable margin in the generation of PDDL domain, achieving over 50% success rate on two tasks (i.e., generating PDDL domains from natural language description or PDDL problems). This is done without requiring additional training. By taking advantage of PDDL as state abstraction, our method is able to outperform current state-of-the-art methods on almost all competition-level planning tasks.

The rapid development of large language model (LLM) alignment algorithms has resulted in a complex and fragmented landscape, with limited clarity on the effectiveness of different methods and their inter-connections. This paper introduces Reward-Aware Preference Optimization (RPO), a mathematical framework that unifies popular preference optimization techniques in LLM alignment, including DPO, IPO, SimPO, and REINFORCE (LOO), among others. RPO provides a structured approach to disentangle and systematically study the impact of various design choices, such as the optimization objective, the number of responses per prompt, and the use of implicit versus explicit reward models, on LLM preference optimization. We additionally propose a new experimental setup that enables the clean and direct ablation of such design choices. Through an extensive series of ablation studies within the RPO framework, we gain insights into the critical factors shaping model alignment, offering practical guidance on the most effective strategies for improving LLM alignment.

One approach to reducing the massive costs of large language models (LLMs) is the use of quantized or sparse representations for training or deployment. While post-training compression methods are very popular, the question of obtaining even more accurate compressed models by directly training over such representations, i.e., Quantization-Aware Training (QAT), is still open: for example, a recent study (arXiv:2411.04330v2) put the "optimal" bit-width at which models can be trained using QAT, while staying accuracy-competitive with standard FP16/BF16 precision, at 8-bits weights and activations. We advance this state-of-the-art via a new method called QuEST, which is Pareto-competitive with FP16, i.e., it provides better accuracy at lower model size, while training models with weights and activations in 4-bits or less. Moreover, QuEST allows stable training with 1-bit weights and activations. QuEST achieves this by improving two key aspects of QAT methods: (1) accurate and fast quantization of the (continuous) distributions of weights and activations via Hadamard normalization and MSE-optimal fitting; (2) a new trust gradient estimator based on the idea of explicitly minimizing the error between the noisy gradient computed over quantized states and the "true" (but unknown) full-precision gradient. Experiments on Llama-type architectures show that QuEST induces stable scaling laws across the entire range of hardware-supported precisions, and can be extended to sparse representations. We provide GPU kernel support showing that models produced by QuEST can be executed efficiently. Our code is available at https://github.com/IST-DASLab/QuEST.

The substantial computational and memory demands of Large Language Models (LLMs) hinder their deployment. Block Floating Point (BFP) has proven effective in accelerating linear operations, a cornerstone of LLM workloads. However, as sequence lengths grow, nonlinear operations, such as Attention, increasingly become performance bottlenecks due to their quadratic computational complexity. These nonlinear operations are predominantly executed using inefficient floating-point formats, which renders the system challenging to optimize software efficiency and hardware overhead. In this paper, we delve into the limitations and potential of applying BFP to nonlinear operations. Given our findings, we introduce a hardware-software co-design framework (DB-Attn), including: (i) DBFP, an advanced BFP version, overcomes nonlinear operation challenges with a pivot-focus strategy for diverse data and an adaptive grouping strategy for flexible exponent sharing. (ii) DH-LUT, a novel lookup table algorithm dedicated to accelerating nonlinear operations with DBFP format. (iii) An RTL-level DBFP-based engine is implemented to support DB-Attn, applicable to FPGA and ASIC. Results show that DB-Attn provides significant performance improvements with negligible accuracy loss, achieving 74% GPU speedup on Softmax of LLaMA and 10x low overhead performance improvement over SOTA designs.