arXiv subjects

Neil Thompson

Publications and source records attributed to Neil Thompson.

28 records · Page 2Linked to original sources

arXiv2025

Introducing the Quantum Economic Advantage Online Calculator

Developing a systematic view of where quantum computers will outperform classical ones is important for researchers, policy makers and business leaders. But developing such a view is challenging because quantum advantage analyses depend not only on algorithm properties, but also on a host of technical characteristics (error correction, gate speeds, etc.). Because various analyses make different assumptions about these technical characteristics, it can be challenging to make comparisons across them. In this paper, we introduce an open-access web-tool designed to make such comparisons easy. Built on the framework introduced by Choi, Moses, and Thompson (2023), it calculates when quantum systems will outperform classical computers for a given algorithmic problem. These estimates can be easily updated based on various assumptions for error correction, overhead, and connectivity. Different hardware roadmaps can also be used and algorithm running times can be customized to particular applications. It can currently be accessed at https://futuretech.mit.edu/quantum-economic-advantage-calculator. This integrated prediction tool also allows us to explore which technical factors are most important for quantum ``economic" advantage (outperforming on a cost-equivalent basis). Overall, we find that for some algorithms (e.g. Shor's) the timing of advantage is quite robust, whereas for others (e.g. Grover's) it is contingent, with numerous technical characteristics substantially impacting these dates. In the paper, we discuss both why this occurs and what we can learn from it.

quant-ph↗

arXiv2025

Meek Models Shall Inherit the Earth

The past decade has seen incredible scaling of AI systems by a few companies, leading to inequality in AI model performance. This paper argues that, contrary to prevailing intuition, the diminishing returns to compute scaling will lead to a convergence of AI model capabilities. In other words, meek models (those with limited computation budget) shall inherit the earth, approaching the performance level of the best models overall. We develop a model illustrating that under a fixed-distribution next-token objective, the marginal capability returns to raw compute shrink substantially. Given current scaling practices, we argue that these diminishing returns are strong enough that even companies that can scale their models exponentially faster than other organizations will eventually have little advantage in capabilities. As part of our argument, we give several reasons that proxies like training loss differences capture important capability measures using evidence from benchmark data and theoretical performance models. In addition, we analyze empirical data on the capability difference of AI models over time. Finally, in light of the increasing ability of meek models, we argue that AI strategy and policy require reexamination, and we outline the areas this shift will affect.

cs.AI↗

arXiv2025

Neural Scaling Laws in Robotics

Neural scaling laws have driven significant advancements in machine learning, particularly in domains like language modeling and computer vision. However, the exploration of neural scaling laws within robotics has remained relatively underexplored, despite the growing adoption of foundation models in this field. This paper represents the first comprehensive study to quantify neural scaling laws for Robot Foundation Models (RFMs) and Large Language Models (LLMs) in robotics tasks. Through a meta-analysis of 327 research papers, we investigate how data size, model size, and compute resources influence downstream performance across a diverse set of robotic tasks. Consistent with previous scaling law research, our results reveal that the performance of robotic models improves with increased resources, following a power-law relationship. Promisingly, the improvement in robotic task performance scales notably faster than language tasks. This suggests that, while performance on downstream robotic tasks today is often moderate-to-poor, increased data and compute are likely to signficantly improve performance in the future. Also consistent with previous scaling law research, we also observe the emergence of new robot capabilities as models scale.

cs.RO↗

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↗

arXiv2024

Intentional and serendipitous diffusion of ideas: Evidence from academic conferences

This paper investigates the effects of seeing ideas presented in-person when they are easily accessible online. Presentations may increase the diffusion of ideas intentionally (when one attends the presentation of an idea of interest) and serendipitously (when one sees other ideas presented in the same session). We measure these effects in the context of 25 computer science conferences using data from the scheduling application Confer, which lets users browse papers, Like those of interest, and receive schedules of their presentations. We address endogeneity concerns in presentation attendance by exploiting scheduling conflicts: when a user Likes multiple papers that are presented at the same time, she cannot see them both, potentially affecting their diffusion. Estimates show that being able to see presentations increases citing of Liked papers within two years by 1.5 percentage points (62.5% boost over the baseline citation rate). Attention to Liked papers also spills over to non-Liked papers in the same session, increasing their citing by 0.5 percentage points (125% boost), and this serendipitous diffusion represents 30.5% of the total effect. Both diffusion types were concentrated among papers semantically close to an attendee's prior work, suggesting that there are inefficiencies in finding related research that conferences help overcome. Overall, even when ideas are easily accessible online, in-person presentations substantially increase diffusion, much of it serendipitous.

cs.DL↗

arXiv2024

The Compute Divide in Machine Learning: A Threat to Academic Contribution and Scrutiny?

There are pronounced differences in the extent to which industrial and academic AI labs use computing resources. We provide a data-driven survey of the role of the compute divide in shaping machine learning research. We show that a compute divide has coincided with a reduced representation of academic-only research teams in compute intensive research topics, especially foundation models. We argue that, academia will likely play a smaller role in advancing the associated techniques, providing critical evaluation and scrutiny, and in the diffusion of such models. Concurrent with this change in research focus, there is a noticeable shift in academic research towards embracing open source, pre-trained models developed within the industry. To address the challenges arising from this trend, especially reduced scrutiny of influential models, we recommend approaches aimed at thoughtfully expanding academic insights. Nationally-sponsored computing infrastructure coupled with open science initiatives could judiciously boost academic compute access, prioritizing research on interpretability, safety and security. Structured access programs and third-party auditing may also allow measured external evaluation of industry systems.

cs.CY↗

arXiv2023

The Quantum Tortoise and the Classical Hare: A simple framework for understanding which problems quantum computing will accelerate (and which it will not)

Quantum computing promises transformational gains for solving some problems, but little to none for others. For anyone hoping to use quantum computers now or in the future, it is important to know which problems will benefit. In this paper, we introduce a framework for answering this question both intuitively and quantitatively. The underlying structure of the framework is a race between quantum and classical computers, where their relative strengths determine when each wins. While classical computers operate faster, quantum computers can sometimes run more efficient algorithms. Whether the speed advantage or the algorithmic advantage dominates determines whether a problem will benefit from quantum computing or not. Our analysis reveals that many problems, particularly those of small to moderate size that can be important for typical businesses, will not benefit from quantum computing. Conversely, larger problems or those with particularly big algorithmic gains will benefit from near-term quantum computing. Since very large algorithmic gains are rare in practice and theorized to be rare even in principle, our analysis suggests that the benefits from quantum computing will flow either to users of these rare cases, or practitioners processing very large data.

cs.DS↗

arXiv2023

Large Language Model Routing with Benchmark Datasets

There is a rapidly growing number of open-source Large Language Models (LLMs) and benchmark datasets to compare them. While some models dominate these benchmarks, no single model typically achieves the best accuracy in all tasks and use cases. In this work, we address the challenge of selecting the best LLM out of a collection of models for new tasks. We propose a new formulation for the problem, in which benchmark datasets are repurposed to learn a "router" model for this LLM selection, and we show that this problem can be reduced to a collection of binary classification tasks. We demonstrate the utility and limitations of learning model routers from various benchmark datasets, where we consistently improve performance upon using any single model for all tasks.

cs.CL↗

arXiv2023

The Grand Illusion: The Myth of Software Portability and Implications for ML Progress

Pushing the boundaries of machine learning often requires exploring different hardware and software combinations. However, the freedom to experiment across different tooling stacks can be at odds with the drive for efficiency, which has produced increasingly specialized AI hardware and incentivized consolidation around a narrow set of ML frameworks. Exploratory research can be restricted if software and hardware are co-evolving, making it even harder to stray away from mainstream ideas that work well with popular tooling stacks. While this friction increasingly impacts the rate of innovation in machine learning, to our knowledge the lack of portability in tooling has not been quantified. In this work, we ask: How portable are popular ML software frameworks? We conduct a large-scale study of the portability of mainstream ML frameworks across different hardware types. Our findings paint an uncomfortable picture -- frameworks can lose more than 40% of their key functions when ported to other hardware. Worse, even when functions are portable, the slowdown in their performance can be extreme and render performance untenable. Collectively, our results reveal how costly straying from a narrow set of hardware-software combinations can be - and suggest that specialization of hardware impedes innovation in machine learning research.

cs.SE↗

arXiv2023

Economic impacts of AI-augmented R&D

Since its emergence around 2010, deep learning has rapidly become the most important technique in Artificial Intelligence (AI), producing an array of scientific firsts in areas as diverse as protein folding, drug discovery, integrated chip design, and weather prediction. As more scientists and engineers adopt deep learning, it is important to consider what effect widespread deployment would have on scientific progress and, ultimately, economic growth. We assess this impact by estimating the idea production function for AI in two computer vision tasks that are considered key test-beds for deep learning and show that AI idea production is notably more capital-intensive than traditional R&D. Because increasing the capital-intensity of R&D accelerates the investments that make scientists and engineers more productive, our work suggests that AI-augmented R&D has the potential to speed up technological change and economic growth.

econ.GN↗