arXiv subjects

Jaime Sevilla

Publications and source records attributed to Jaime Sevilla.

13 recordsLinked to original sources

arXiv2026

Frontier AI Auditing: Toward Rigorous Third-Party Assessment of Safety and Security Practices at Leading AI Companies

We outline a vision for frontier AI auditing, which we define as rigorous third-party verification of frontier AI developers' safety and security claims, and evaluation of their systems and practices against relevant standards, based on deep, secure access to non-public information. Frontier AI audits should not be limited to a company's publicly deployed products, but should instead consider the full range of organization-level safety and security risks, including internal deployment of AI systems, information security practices, and safety decision-making processes. We describe four AI Assurance Levels (AALs), the higher levels of which provide greater confidence in audit findings. We recommend AAL-1 as a baseline for frontier AI generally, and AAL-2 as a near-term goal for the most advanced subset of frontier AI developers. Achieving the vision we outline will require (1) ensuring high quality standards for frontier AI auditing, so it does not devolve into a checkbox exercise or lag behind changes in the industry; (2) growing the ecosystem of audit providers at a rapid pace without compromising quality; (3) accelerating adoption of frontier AI auditing by clarifying and strengthening incentives; and (4) achieving technical readiness for high AI Assurance Levels so they can be applied when needed.

cs.CY↗

arXiv2025

GATE: An Integrated Assessment Model for AI Automation

Assessing the economic impacts of artificial intelligence requires integrating insights from both computer science and economics. We present the Growth and AI Transition Endogenous model (GATE), a dynamic integrated assessment model that simulates the economic effects of AI automation. GATE combines three key ingredients that have not been brought together in previous work: (1) a compute-based model of AI development, (2) an AI automation framework, and (3) a semi-endogenous growth model featuring endogenous investment and adjustment costs. The model allows users to simulate the economic effects of the transition to advanced AI across a range of potential scenarios. GATE captures the interactions between economic variables, including investment, automation, innovation, and growth, as well as AI-related inputs such as compute and algorithms. This paper explains the model's structure and functionality, emphasizing AI development for economists and economic modeling for the AI community. The model is implemented in an interactive sandbox, enabling users to explore the impact of AI under different parameter choices and policy interventions. The modeling sandbox is available at: www.epoch.ai/GATE.

econ.GN↗

arXiv2024

FrontierMath: A Benchmark for Evaluating Advanced Mathematical Reasoning in AI

We introduce FrontierMath, a benchmark of hundreds of original, exceptionally challenging mathematics problems crafted and vetted by expert mathematicians. The questions cover most major branches of modern mathematics -- from computationally intensive problems in number theory and real analysis to abstract questions in algebraic geometry and category theory. Solving a typical problem requires multiple hours of effort from a researcher in the relevant branch of mathematics, and for the upper end questions, multiple days. FrontierMath uses new, unpublished problems and automated verification to reliably evaluate models while minimizing risk of data contamination. Current state-of-the-art AI models solve under 2% of problems, revealing a vast gap between AI capabilities and the prowess of the mathematical community. As AI systems advance toward expert-level mathematical abilities, FrontierMath offers a rigorous testbed that quantifies their progress.

cs.AI↗

arXiv2024

Explaining Bayesian Networks in Natural Language using Factor Arguments. Evaluation in the medical domain

In this paper, we propose a model for building natural language explanations for Bayesian Network Reasoning in terms of factor arguments, which are argumentation graphs of flowing evidence, relating the observed evidence to a target variable we want to learn about. We introduce the notion of factor argument independence to address the outstanding question of defining when arguments should be presented jointly or separately and present an algorithm that, starting from the evidence nodes and a target node, produces a list of all independent factor arguments ordered by their strength. Finally, we implemented a scheme to build natural language explanations of Bayesian Reasoning using this approach. Our proposal has been validated in the medical domain through a human-driven evaluation study where we compare the Bayesian Network Reasoning explanations obtained using factor arguments with an alternative explanation method. Evaluation results indicate that our proposed explanation approach is deemed by users as significantly more useful for understanding Bayesian Network Reasoning than another existing explanation method it is compared to.

cs.AI↗

arXiv2024

Will we run out of data? Limits of LLM scaling based on human-generated data

We investigate the potential constraints on LLM scaling posed by the availability of public human-generated text data. We forecast the growing demand for training data based on current trends and estimate the total stock of public human text data. Our findings indicate that if current LLM development trends continue, models will be trained on datasets roughly equal in size to the available stock of public human text data between 2026 and 2032, or slightly earlier if models are overtrained. We explore how progress in language modeling can continue when human-generated text datasets cannot be scaled any further. We argue that synthetic data generation, transfer learning from data-rich domains, and data efficiency improvements might support further progress.

cs.LG↗

arXiv2024

Algorithmic progress in language models

We investigate the rate at which algorithms for pre-training language models have improved since the advent of deep learning. Using a dataset of over 200 language model evaluations on Wikitext and Penn Treebank spanning 2012-2023, we find that the compute required to reach a set performance threshold has halved approximately every 8 months, with a 95% confidence interval of around 5 to 14 months, substantially faster than hardware gains per Moore's Law. We estimate augmented scaling laws, which enable us to quantify algorithmic progress and determine the relative contributions of scaling models versus innovations in training algorithms. Despite the rapid pace of algorithmic progress and the development of new architectures such as the transformer, our analysis reveals that the increase in compute made an even larger contribution to overall performance improvements over this time period. Though limited by noisy benchmark data, our analysis quantifies the rapid progress in language modeling, shedding light on the relative contributions from compute and algorithms.

cs.CL↗

arXiv2023

Power Law Trends in Speedrunning and Machine Learning

We find that improvements in speedrunning world records follow a power law pattern. Using this observation, we answer an outstanding question from previous work: How do we improve on the baseline of predicting no improvement when forecasting speedrunning world records out to some time horizon, such as one month? Using a random effects model, we improve on this baseline for relative mean square error made on predicting out-of-sample world record improvements as the comparison metric at a $p < 10^{-5}$ significance level. The same set-up improves \textit{even} on the ex-post best exponential moving average forecasts at a $p = 0.15$ significance level while having access to substantially fewer data points. We demonstrate the effectiveness of this approach by applying it to Machine Learning benchmarks and achieving forecasts that exceed a baseline. Finally, we interpret the resulting model to suggest that 1) ML benchmarks are far from saturation and 2) sudden large improvements in Machine Learning are unlikely but cannot be ruled out.

cs.LG↗

arXiv2022

Machine Learning Model Sizes and the Parameter Gap

We study trends in model size of notable machine learning systems over time using a curated dataset. From 1950 to 2018, model size in language models increased steadily by seven orders of magnitude. The trend then accelerated, with model size increasing by another five orders of magnitude in just 4 years from 2018 to 2022. Vision models grew at a more constant pace, totaling 7 orders of magnitude of growth between 1950 and 2022. We also identify that, since 2020, there have been many language models below 20B parameters, many models above 70B parameters, but a scarcity of models in the 20-70B parameter range. We refer to that scarcity as the parameter gap. We provide some stylized facts about the parameter gap and propose a few hypotheses to explain it. The explanations we favor are: (a) increasing model size beyond 20B parameters requires adopting different parallelism techniques, which makes mid-sized models less cost-effective, (b) GPT-3 was one order of magnitude larger than previous language models, and researchers afterwards primarily experimented with bigger models to outperform it. While these dynamics likely exist, and we believe they play some role in generating the gap, we don't have high confidence that there are no other, more important dynamics at play.

cs.LG↗

arXiv2022

Compute Trends Across Three Eras of Machine Learning

Compute, data, and algorithmic advances are the three fundamental factors that guide the progress of modern Machine Learning (ML). In this paper we study trends in the most readily quantified factor - compute. We show that before 2010 training compute grew in line with Moore's law, doubling roughly every 20 months. Since the advent of Deep Learning in the early 2010s, the scaling of training compute has accelerated, doubling approximately every 6 months. In late 2015, a new trend emerged as firms developed large-scale ML models with 10 to 100-fold larger requirements in training compute. Based on these observations we split the history of compute in ML into three eras: the Pre Deep Learning Era, the Deep Learning Era and the Large-Scale Era. Overall, our work highlights the fast-growing compute requirements for training advanced ML systems.

cs.LG↗

arXiv2021

Finding, Scoring and Explaining Arguments in Bayesian Networks

We propose a new approach to explain Bayesian Networks. The approach revolves around a new definition of a probabilistic argument and the evidence it provides. We define a notion of independent arguments, and propose an algorithm to extract a list of relevant, independent arguments given a Bayesian Network, a target node and a set of observations. To demonstrate the relevance of the arguments, we show how we can use the extracted arguments to approximate message passing. Finally, we show a simple scheme to explain the arguments in natural language.

cs.AI↗

arXiv2021

A conditional independence test for causality in econometrics

The Y-test is a useful tool for detecting missing confounders in the context of a multivariate regression.However, it is rarely used in practice since it requires identifying multiple conditionally independent instruments, which is often impossible. We propose a heuristic test which relaxes the independence requirement. We then show how to apply this heuristic test on a price-demand and a firm loan-productivity problem. We conclude that the test is informative when the variables are linearly related with Gaussian additive noise, but it can be misleading in other contexts. Still, we believe that the test can be a useful concept for falsifying a proposed control set.

stat.ME↗

arXiv2020

Forecasting timelines of quantum computing

We consider how to forecast progress in the domain of quantum computing. For this purpose we collect a dataset of quantum computer systems to date, scored on their physical qubits and gate error rate, and we define an index combining both metrics, the generalized logical qubit. We study the relationship between physical qubits and gate error rate, and tentatively conclude that they are positively correlated (albeit with some room for doubt), indicating a frontier of development that trades-off between them. We also apply a log-linear regression on the metrics to provide a tentative upper bound on how much progress can be expected over time. Within the (generally optimistic) assumptions of our model, including the key assumption that exponential progress in qubit count and gate fidelity will continue, we estimate that that proof-of-concept fault-tolerant computation based on superconductor technology is unlikely (<5% confidence) to be exhibited before 2026, and that quantum devices capable of factoring RSA-2048 are unlikely (<5% confidence) to exist before 2039. It is of course possible that these milestones will in fact be reached earlier, but that this would require faster progress than has yet been seen.

quant-ph↗

arXiv2019

Implications of Quantum Computing for Artificial Intelligence alignment research

We explain some key features of quantum computing via three heuristics and apply them to argue that a deep understanding of quantum computing is unlikely to be helpful to address current bottlenecks in Artificial Intelligence Alignment. Our argument relies on the claims that Quantum Computing leads to compute overhang instead of algorithmic overhang, and that the difficulties associated with the measurement of quantum states do not invalidate any major assumptions of current Artificial Intelligence Alignment research agendas. We also discuss tripwiring, adversarial blinding, informed oversight and side effects as possible exceptions.

cs.ET↗