Machine Learning at the Patent Office: Lessons for Patents and Administrative Law

• Author: Arti K. Rai
• Position: Elvin R. Latty Professor and Director, Center for Innovation Policy, Duke Law School
• Pages: 2617-2641

2617

Machine Learning at the Patent Office:

Lessons for Patents and

Administrative Law

Arti K. Rai*

ABSTRACT: The empirical data indicate that a relatively small increment

of additional U.S. Patent and Trademark Office (“Patent Office” or

“USPTO”) investment in prior art search at the initial examination stage

could be a cost-effective mechanism for improving accuracy in the patent

system. This contribution argues that machine learning provides a promising

arena for such investment. Notably, the use of machine learning in patent

examination does not raise the same potent concerns about individual rights

and discrimination that it raises in other areas of administrative and judicial

process. To be sure, even an apparently easy case like prior art search at the

USPTO poses challenges. The most important generalizable challenge relates

to explainability. The USPTO has stressed transparency to the general public

as necessary for achieving adequate explainability. However, at least in

contexts like prior art search, adequate explainability does not require full

transparency. Moreover, full transparency would chill provision of private

sector expertise and would be susceptible to gaming.

I. INTRODUCTION ........................................................................... 2618

II.THE CASE FOR MORE INTENSIVE EX ANTE EXAMINATION ........... 2621

III. AN EASY CASE? ............................................................................ 2624

A.EXPLAINABILITY: GENERAL CONSIDERATIONS ......................... 2625

B.EXPLAINABILITY IN PRIOR ART SEARCH.................................. 2629

IV.THE PTO AND MACHINE LEARNING: COMPARING THEORY

AND PRACTICE ............................................................................. 2632

*

Elvin R. Latty Professor and Director, Center for Innovation Policy, Duke Law School.

I presented earlier versions of these ideas at the University of Iowa College of Law symposiu m on

“Administering Patent Law,” the 18th Annual Intellectual Property Scholars Conference, and a

May 2018 Duke Law symposium on “AI in the Administrative State.” I thank the participants at

those fora, particularly Scott Beliveau, Stuart Benjamin, Cary Coglianese, Ian Weatherbee, and

Ryan Whalen for very helpful comments. Bennett Wright provided superb research assistance.

2618 IOWA LAW REVIEW [Vol. 104:2617

A.THE BASICS OF SEARCH ......................................................... 2632

B.STEPS TOWARD MACHINE LEARNING...................................... 2634

C.CHALLENGES OF HUMAN CAPITAL AND DATA ......................... 2636

D.EXPLAINABILITY AND TRANSPARENCY CHALLENGES ................ 2638

V.CONCLUSION .............................................................................. 2640

I. INTRODUCTION

A voluminous body of scholarly and popular commentary discusses the

use of predictive algorithms by government actors. The decision rule in

question can be explicitly specified by humans. Conventional linear

regression, for example, is a specific, human-generated data model that

transforms inputs into outputs.1 Alternatively, the decision rule can emerge

from algorithmic or machine learning. Although machine learning

encompasses many algorithms of varying complexity, a distinctive feature of

the genre is that the learning algorithm does not represent the decision rule;

instead, the algorithm “learns” the decision rules from data known as training

data.2

In both cases, the commentary has often been highly critical, particularly

in addressing deployment of algorithms in areas like predictive policing and

criminal risk assessment.3 Commentators have expressed concern about

classification based on legally protected characteristics and inaccurate

adjudication of individual rights.4

Understandably, legal commentators have paid less attention to decision-

making contexts where bias and rights are not first-order concerns.5 Yet those

contexts, which can involve decisions that are quite important for social

welfare, are also worthy of study.

1. RICHARD A. BERK, STATISTICAL LEARNING FROM A REGRESSION PERSPECTIVE 327 (2d ed. 2016).

2. See generally David Lehr & Paul Ohm, Playing with the Data: What Legal Scholars Should Learn

About Machine Learning, 51 U.C. DAVIS L. REV. 653 (2017) (providing a review of machine learning

and drawing attention to ways in which legal scholars have mischaracterized machine learning).

3. See, e.g., Elizabeth E. Joh, The New Surveillance Discretion: Au tomated Suspicion, Big Data,

and Policing, 10 HARV. L. & POL’Y REV. 15, 15–19, 30–32 (2016); Dawinder S. Sidhu, Moneyball

Sentencing, 56 B.C. L. REV. 671, 673–76 (2015); Sonja B. Starr, Evidence-Based Sentencing and the

Scientific Rationalization of Discrimination, 66 STAN. L. REV. 803, 805–08 (2014). In the area of

criminal justice, a somewhat related literature addresses the issue of forensic evidence generated

by software. See generally Andrea Roth, Machine Testimony, 126 YALE L.J. 1972 (2017) (arguing that

such testimony presents challenges).

4. See sources cited supra note 3.

5. For a notable exception, see generally Cary Coglianese & David Lehr, Regulating by Robot:

Administrative Decision Making in the Machine-Learning Era, 105 GEO. L.J. 1147 (2017) (discussing

a broad range of administrative decision making).