Scaling law that rewards large datasets, complex models and enhanced data granularity has been observed in various fields of deep learning. Yet, studies on time series forecasting have cast doubt on scaling behaviors of deep learning methods for time series forecasting: while more training data improves performance, more capable models do not always outperform less capable models, and longer input horizon may hurt performance for some models. We propose a theory for scaling law for time series forecasting that can explain these seemingly abnormal behaviors. We take into account the impact of dataset size and model complexity, as well as time series data granularity, particularly focusing on the look-back horizon, an aspect that has been unexplored in previous theories. Furthermore, we empirically evaluate various models using a diverse set of time series forecasting datasets, which (1) verifies the validity of scaling law on dataset size and model complexity within the realm of time series forecasting, and (2) validates our theoretical framework, particularly regarding the influence of look back horizon. We hope our findings may inspire new models targeting time series forecasting datasets of limited size, as well as large foundational datasets and models for time series forecasting in future works.

Scaling law that rewards large datasets, complex models and enhanced data granularity has been observed in various fields of deep learning. Yet, studies on time series forecasting have cast doubt on scaling behaviors of deep learning methods for time series forecasting: while more training data improves performance, more capable models do not always outperform less capable models, and longer input horizon may hurt performance for some models. We propose a theory for scaling law for time series forecasting that can explain these seemingly abnormal behaviors. We take into account the impact of dataset size and model complexity, as well as time series data granularity, particularly focusing on the look-back horizon, an aspect that has been unexplored in previous theories. Furthermore, we empirically evaluate various models using a diverse set of time series forecasting datasets, which (1) verifies the validity of scaling law on dataset size and model complexity within the realm of time series forecasting, and (2) validates our theoretical framework, particularly regarding the influence of look back horizon. We hope our findings may inspire new models targeting time series forecasting datasets of limited size, as well as large foundational datasets and models for time series forecasting in future works.

Earlier this week via blog post, OpenAI released their newest AI models: o3 and o4-mini. These models are the company's "smartest and most capable models to date" and their first reasoning models that can also reason when it comes to images. In short, these AI models can use an image--such as a photograph or a sketch--as part of an analysis. The models can also adjust, zoom in on, and rotate an image during reasoning. For the first time, our reasoning models can agentically use and combine every tool within ChatGPT, including web search, Python, image analysis, file interpretation, and image generation.

Zico Kolter has a knack for getting artificial intelligence to misbehave in interesting and important ways. His research group at Carnegie Mellon University has discovered numerous methods of tricking, goading, and confusing advanced AI models into being their worst selves. Kolter is a professor at CMU, a technical adviser to Gray Swan, a startup specializing in AI security, and, as of August 2024, a board member at the world's most prominent AI company, OpenAI. In addition to pioneering ways of jailbreaking commercial AI models, Kolter designs his own models that are more secure by nature. As AI becomes more autonomous, Kolter believes that AI agents may pose unique challenges--especially when they start talking to one another.

The trustworthiness of highly capable language models is put at risk when they are able to produce deceptive outputs. Moreover, when models are vulnerable to deception it undermines reliability. In this paper, we introduce a method to investigate complex, model-on-model deceptive scenarios. We create a dataset of over 10,000 misleading explanations by asking Llama-2 7B, 13B, 70B, and GPT-3.5 to justify the wrong answer for questions in the MMLU. We find that, when models read these explanations, they are all significantly deceived. Worryingly, models of all capabilities are successful at misleading others, while more capable models are only slightly better at resisting deception. We recommend the development of techniques to detect and defend against deception. Since the release of OpenAI's ChatGPT, large language models (LLMs) have revolutionized information accessibility by providing precise answers and supportive explanations to complex queries (Spatharioti et al., 2023; Caramancion, 2024; OpenAI, 2022). However, LLMs have also demonstrated a propensity to hallucinate explanations that are convincing but incorrect (Zhang et al., 2023; Walters & Wilder, 2023; Xu et al., 2024).

Azure OpenAI provides access to many different models, grouped by family and capability. A model family typically associates models by their intended task. The following table describes model families currently available in Azure OpenAI. Not all models are available in all regions currently. Each model family has a series of models that are further distinguished by capability.

People haven't exactly been impressed in the short time since Google released its "experimental conversational AI service" Bard. Coming up against OpenAI's ChatGPT and Microsoft's Bing Chat (also powered by OpenAI's GPT-4) users have found its responses to not be as knowledgeable or detailed as its rivals. That could be set to change, however, after Google CEO Sundar Pichai confirmed on The New York Times podcast "Hard Fork" that Bard will soon be moving from its current LaMDA-based model to larger-scale PaLM datasets in the coming days. When asked how he felt about responses to Bard's release, Pichai commented: "We clearly have more capable models. Pretty soon, maybe as this goes live, we will be upgrading Bard to some of our more capable PaLM models, so which will bring more capabilities, be it in reasoning, coding." To frame the difference, Google said it had trained LaMDA with 137 billion parameters when it shared details about the language-based models last year.