Peter Norvig

Peter Norvig
Google Inc. | Google · Research Department

Doctor of Philosophy

About

132
Publications
67,258
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
44,112
Citations
Introduction
Skills and Expertise

Publications

Publications (132)
Article
The growing importance of machine learning creates challenging questions for computing education.
Chapter
At the Brains, Minds, and Machines symposium held during MIT’s 150th birthday party, Technology Review reports that Prof. Noam Chomsky derided researchers in machine learning who use purely statistical methods to produce behavior that mimics something in the world, but who don’t try to understand the meaning of that behavior. This essay discusses w...
Article
Full-text available
Neural networks that compute over graph structures are a natural fit for problems in a variety of domains, including natural language (parse trees) and cheminformatics (molecular graphs). However, since the computation graph has a different shape and size for every input, such networks do not directly support batched training or inference. They are...
Conference Paper
We would like to understand the meaning of content on the web. Bit where should that meaning come from? From markup language annotations created by the authors of the content? Crowdsourced from readers of the content? Automatically extracted by machine learning algorithms? This talk investigates the possibilities.
Conference Paper
Certain tasks, such as recognizing speech, or correcting spelling errors, are now routinely handled with machine learning algorithms. But most tasks are handled the old fashioned way, with programmers writing code line by line. Machine learning algorithms work by amassing large numbers of examples and extracting patterns from them. We certainly hav...
Conference Paper
There is great enthusiasm for the idea that massive amounts of data from online interactions of learners with material can lead to a rapid improvement cycle, driven by analysis of the data, experimentation, and intervention to do more of what works and less of what doesn't. This talk discusses techniques for working with massive amounts of data. Pe...
Conference Paper
Today, artificial intelligence systems approve credit card transactions, prevent fraud, trade stocks, recognize faces, discover patterns in genomic data, and help doctors interpret test results. In the future, how can we build reliable systems from components that learn, adapt, interact, and change? Are progress and reliability fundamentally incomp...
Conference Paper
If you want to recognize speech or filter out spam emails, you will probably write a machine learning algorithm and will not try to write the whole program using a "traditional" software specification and implementation. There are many examples of successful machine learning solutions, but can we more broadly apply the techniques to most or all sof...
Conference Paper
The IPython Notebook is an interactive browser-based environment where you can combine code execution, text, mathematics, plots, and rich media into a single document. Originally designed for use as an electronic lab notebook for computational science, it is increasingly being used in teaching as well, and a rich ecosystem of open source plugins an...
Conference Paper
The software platforms that mediate online learning experiences are the common ground where learning science and computer science intersect. This panel will discuss the affordances of current online learning platforms and lessons learned in using them with students. The goal of the panel is to help learning scientists and computer scientists unders...
Patent
Full-text available
A rewrite component automatically generates rewrite rules that describe how uniform resource locators (URLs) can be rewritten to reduce or eliminate different URLs that redundantly refer to the same or substantially the same content. The rewrite rules can be applied to URLs received when crawling a network to increase the efficiency of the crawl an...
Article
Full-text available
How thousands of online students can get the effect of one-on-one tutoring
Article
Full-text available
We have long imagined machines that can reason and learn as well as a human can. But it has turned out to be surprisingly difficult
Article
AIs route phone calls, approve credit card transactions, prevent fraud, trade stocks, recognise faces, and even help doctors interpret test results
Article
Neuroscience and psychology are leading to new hardware and software designs, and the internet provides a vast store of data. Will AI evolve on its own?
Article
As our tools get ever better, we will become better as well. We may hear less about AI and more about IA – “intelligence amplification”
Article
Language recognition programs use massive databases of words, and statistical correlations between those words, to translate or to recognise speech. But correlation is not causation. Do these statistical data‐dredgings give any insight into how language works? Or are they a mere big‐number trick, useful but adding nothing to understanding? One who...
Article
Full-text available
By closely connecting research and development Google is able to conduct experiments on an unprecedented scale, often resulting in new capabilities for the company.
Conference Paper
Full-text available
This talk covers techniques for analyzing data sets with up to trillions of examples with billions of features, using thousands of computers. To operate at this scale requires an understanding of an increasing complex hardware hierarchy (e.g. cache, memory, SSD, another machine in the rack, disk, a machine in another data center, ...); a model for...
Article
Full-text available
We constructed a corpus of digitized texts containing about 4% of all books ever printed. Analysis of this corpus enables us to investigate cultural trends quantitatively. We survey the vast terrain of ‘culturomics,’ focusing on linguistic and cultural phenomena that were reflected in the English language between 1800 and 2000. We show how this app...
Article
Full-text available
For the first issue of the new decade, Nature asked a selection of leading researchers and policy-makers where their fields will be ten years from now. We invited them to identify the key questions their disciplines face, the major roadblocks and the pressing next steps.
Article
Full-text available
At Brown University, there is excitement of having access to the Brown Corpus, containing one million English words. Since then, we have seen several notable corpora that are about 100 times larger, and in 2006, Google released a trillion-word corpus with frequency counts for all sequences up to five words long. In some ways this corpus is a step b...
Conference Paper
Full-text available
Growing email volumes cause flooded inboxes and swelled email archives, making search and new email processing difficult. While emails have rich metadata, such as recipients and folders, suitable for creating filtered views, it is often difficult to choose appropriate filters for new inbox messages without first examining messages. In this work, we...
Conference Paper
Full-text available
The problem of deciding what was implied by a written text, of "reading between the lines" is the problem of inference. To extract proper inferences from a text requires a great deal of general knowledge on the part of the reader. Past approaches have often postulated an algorithm tuned to process a particular kind of knowledge structure (such as a...
Article
Various disciplines have examined the many phenomena of metacognition and have produced numerous results, both positive and negative. I discuss some of these aspects of cognition about cognition and the results concerning them from the point of view ...
Article
Full-text available
This paper benefined from discussions with Michael Braveman, Dan Jurafsky, Nigel Ward, Dekai Wu, and other members of the BAIR seminar
Article
Full-text available
This paper describes an integrated neural flight and propulsion control system. which uses a neural network based approach for applying alternate sources of control power in the presence of damage or failures. Under normal operating conditions, the system utilizes conventional flight control surfaces. Neural networks are used to provide consistent...
Article
The title of this book is probably misleading in a number of ways. A potential UNIX user might pick it up hoping to make a decision about what software product to acquire. Even if it is properly placed in the artificial intelligence (AI) section of the book store, it is likely to be taken as yet another set of case studies of AI applications to a p...
Article
Full-text available
The NASA's Intelligent Synthesis Environment (ISE) program is a grand attempt to develop a system to transform the way complex artifacts are engineered. This paper discusses a "middleware" architecture for enabling the development of ISE. Desirable elements of such an Intelligent Synthesis Architecture (ISA) include remote invocation; plug-and-play...
Article
This is the last of a series of three special issues on intelligent help systems for UNIX.1 This issue addresses natural language dialogue whereas the previous issues focussed on computational models and systems and planning and knowledge representation, respectively. In this collection, the focus is not upon issues of parsing and production per se...
Article
We have realized the help system SINIX Consultant (SC) for SINIX1 users. The system is capable of answering – in German – natural language questions concerning SINIX commands, objects, and concepts. But not only does this ...
Article
This is the first in a series of three special issues focussed on intelligent help systems for UNIX, each with its own emphasis: (1) computational models and systems, (2) planning and knowledge representation and (3) natural language dialogue. In this first issue focussing on computational models and systems there are five papers, one addressing em...
Book
In this international collection of papers there is a wealth of knowledge on artificial intelligence (AI) and cognitive science (CS) techniques applied to the problem of providing help systems mainly for the UNIX operating system. The research described here involves the representation of technical computer concepts, but also the representation of...
Article
Full-text available
In this paper we argue that LISP can provide a powerful tool for web programmers, at both the novice and expert level, and that LISP has the potential to become a language of choice for writing both client-side and server-side internet programs. The syntactic and semantic simplicity of LISP enables non-experts to quickly master a basic level of LIS...
Article
This presentation gives a summary of intelligent agents for design synthesis environments. We'll start with the conclusions, and work backwards to justify them. First, an important assumption is that agents (whatever they are) are good for software engineering. This is especially true for software that operates in an uncertain, changing environment...
Article
This paper describes an Integrated Vehicle Modeling Environment for estimating aircraft geometric, inertial, and aerodynamic characteristics, and for interfacing with a high fidelity, workstation based flight simulation architecture. The goals in developing this environment are to aid in the design of next generation intelligent fight control techn...
Article
Much of the world's data lies outside relational databases-scattered across Web sites, file systems, nonrelational databases, and legacy applications. These data sources differ in the way they organize data, in the vocabulary they use, and in their data-access or query mechanisms. These differences make it difficult to combine data from multiple so...
Article
In the commercial world, the biggest hurdles to delivering embedded intelligence in software are the software-engineering costs. We members of the intelligent-systems community have brilliant techniques available, and no end of applications is in sight, but the gap between intelligent-systems techniques and delivered software is too great. Standard...
Article
In this talk I will present some video-materials gathered in Mission Control during simulations. The focus of the presentation will be on verbal and non-verbal communication between the officers in the front and backroom, especially the practices that have evolved around a peculiar communications technology called voice loops.
Article
stitution at the end of the algorithm. This approach does not generate intermediate garbage, but it may take more time, because it is necessary to search the substitution list for the values of variables. The trade-off is justified when patterns are large relative to the number of variables in them. It is not just textbooks that follow this model;...
Article
Many people see agents and agent-based programming as ushering in a new era in computing - particularly on the Internet. This paper explains why programming agents is not just business as usual, but requires a new way of looking at problems and their solutions. It also examines how reinforcement learning can be used to address these problems.
Article
We describe our experiences in teaching introductory AI and in writing a textbook for the course. The book tries to make the concepts of AI more concrete via two strategies: relating them to the student's existing knowledge, and using examples based on an agent operating in an environment.
Chapter
The long-anticipated revision of this best-selling book offers the most comprehensive, up-to-date introduction to the theory and practice of artificial intelligence. Intelligent Agents. Solving Problems by Searching. Informed Search Methods. Game Playing. Agents that Reason Logically. First-order Logic. Building a Knowledge Base. Inference in First...
Article
Full-text available
Knowledge Interchange Format (KIF) is a computer-oriented language for the interchange of knowledge among disparate programs. It has declarative semantics (i.e. the meaning of expressions in the representation can be understood without appeal to an interpreter for manipulating those expressions); it is logically comprehensive (i.e. it provides for...
Article
Contenido: I. Introducción a Common Lisp: 1. Introducción a Lisp; 2. Un programa en Lisp sencillo; 3. Revisión de Lisp; II. Programas elementales de inteligencia artificial: 4. GPS: el solucionador general de programas; 5. ELIZA: diálogo con la máquina; 6. Construcción de herramientas de software; 7. STUDENT: solución de problemas de palabras en ál...
Article
Full-text available
A queue is a data structure where items are entered one at a time and removed one at a time in the same order---i.e., first in first out. They are the same as stacks except that in a stack, items are removed in the reverse of the order they are entered---i.e., last in first out. Queues are most precisely described by the functions that act on them:...
Article
The unification of two patterns both containing variables is a ubiquitous operation in logic programming and in many artificial intelligence applications. Thus, many texts present unification algorithms. Unfortunately, at least seven of these presentations are incorrect. The common error occurs when logic variables are represented as binding lists;...