Relatively Speaking: Predictive Coding Technologies and Protocols Survey Results

The Predictive Coding Technologies and Protocols Survey is a non-scientific survey designed to help provide a general understanding of the use of predictive coding technologies and protocols from data discovery and legal discovery professionals within the eDiscovery ecosystem.

Article by Rob Robinson

The Predictive Coding Technologies and Protocols Survey

Initiated on August 31, 2018, the Predictive Coding Technologies and Protocols Survey is a non-scientific survey designed to help provide a general understanding of the use of predictive coding technologies and protocols from data discovery and legal discovery professionals within the eDiscovery ecosystem. The survey was open from August 31 through September 15, 2018, with individuals invited to participate directly by ComplexDiscovery and indirectly by industry website, blog, and newsletter mentions.

Designed to provide a general understanding of predictive coding technologies and protocols, the survey had two primary educational objectives:

  • To provide a consolidated listing of potential predictive coding technology and protocol definitions. While not all-inclusive or comprehensive, the listing was vetted with selected industry predictive coding experts for completeness and accuracy, thus it appears to be profitable for use in educational efforts.
  • To ask eDiscovery ecosystem professionals about their usage and preferences of predictive coding platforms, technologies, and protocols.

The survey offered responders an opportunity to provide predictive coding background information, including their primary predictive coding platform, as well as posed three specific questions to responders. Those questions being:

  • Which predictive coding technologies are utilized by your eDiscovery platform?
  • If you use k-Nearest Neighbor Classifier (k-NN), do you use 1-NN or k-NN with k>1?
  • Which technology-assisted review protocols are utilized in your delivery of predictive coding?

Closed on September 15, 2018, the survey had 31 responders.

Key Results and Observations

Primary Predictive Coding Platform (Chart 1)

  • More than 80% of responders (80.64%) shared that they did have a specific primary platform for predictive coding.
  • There were 12 different platforms noted as primary predictive platforms by responders.
  • Relativity was listed as the primary predictive coding platform by more than 30% of responders (32.25%).
  • The top three survey primary predictive coding platforms accounted for more than 50% of responses (51.61%).
  • Just under 20% of responders (19.35%) shared they had no single primary platform for predictive coding.

Primary Predictive Coding Technologies Used (Chart 2)

  • All listed predictive coding technologies plus one added technology are used by at least one survey responder.
  • Active Learning was the most used predictive coding technology, with more than 70% of responders (70.96%) reporting that they use it in their predictive coding efforts.
  • Approximately 32% of responders (32.25%) only use one predictive coding technology in their predictive coding efforts.
  • Approximately 68% of responders (67.74%) use more than one predictive coding technology in their predictive coding efforts.

Usage of k-Nearest Neighbor Classifier (Charts 3 and 4)

  • Approximately 25% of responders (25.80%) reported use of k-Nearest Neighbor Classifier as a predictive coding technology.
  • Of those responders using k-Nearest Neighbor Classifier as a predictive coding technology, 50% use 1-NN, and 50% use k-NN with k>1.

Technology-Assisted Review Protocols for Predictive Coding Used (Chart 5)

  • All listed technology-assisted protocols for predictive coding were used by at least one survey responder.
  • Continuous Active Learning (CAL) was the most used predictive coding protocol, with more than 87% of responders (87.09%) reporting that they use it in their predictive coding efforts.
  • Approximately 50% of responders (48.38%) use only one predictive coding protocol in their predictive coding efforts.
  • Approximately 50% of responders (51.61%) use more than one predictive coding protocol in their predictive coding efforts.

Predictive Coding Technology and Protocol Survey Responder Overview (Chart 6)

  • Approximately 54% of responders (54.83%) are from technology provider organizations.
  • Approximately 32% of responders (32.25%) are from law firms.

Survey Charts

(Charts can be expanded for detailed viewing.)

Chart 1: Name of Primary Predictive Coding Platform

Primary Predictive Coding Platform (31 Responders)

Chart 2: Which predictive coding technologies are utilized by your eDiscovery platform?

Predictive Coding Technologies Used (31 Responders)

Chart 3: Usage of k-Nearest Neighbor Classifier (k-NN)

Usage of k-Nearest Neighbor Classifier (31 Responders)

Chart 4: If you use k-Nearest Neighbor Classifier (k-NN), do you use 1-NN or k-NN with k>1?

Usage of k-Nearest Neighbor Classifier (k-NN) (8 Responders)

Chart 5: Which technology-assisted review protocols are utilized in your delivery of predictive coding?

Predictive Coding Protocols Used (31 Responders)

Chart 6: Survey Responder Overview

Responder Overview (31 Responders)

Survey Background Information

A Working List of Predictive Coding Technologies (1,2,3,4)

Aggregated from electronic discovery experts based on professional publications and personal conversations, provided below is a non-all inclusive working list of identified machine learning technologies that have been applied or have the potential to be applied to the discipline of eDiscovery to facilitate predictive coding. This working list is designed to provide a reference point for identified predictive coding technologies and may over time include additions, adjustments, and amendments based on feedback from experts and organizations applying and implementing these mainstream technologies in their specific eDiscovery platforms.

Listed in Alphabetical Order

  • Active Learning: A process, typically iterative, whereby an algorithm is used to select documents that should be reviewed for training based on a strategy to help the classification algorithm learn efficiently.
  • Decision Tree: A step-by-step method of distinguishing between relevant and non-relevant documents, depending on what combination of words (or other features) they contain. A Decision Tree to identify documents pertaining to financial derivatives might first determine whether or not a document contained the word “swap.” If it did, the Decision Tree might then determine whether or not the document contained “credit,” and so on. A Decision Tree may be created either through knowledge engineering or machine learning.
  • k-Nearest Neighbor Classifier (k-NN): A classification algorithm that analyzes the k example documents that are most similar (nearest) to the document being classified in order to determine the best classification for the document. If k is too small (e.g., k=1), it may be extremely difficult to achieve high recall.
  • Latent Semantic Analysis (LSA): A mathematical representation of documents that treats highly correlated words (i.e., words that tend to occur in the same documents) as being, in a sense, equivalent or interchangeable. This equivalency or interchangeability can allow algorithms to identify documents as being conceptually similar even when they aren’t using the same words (e.g., because synonyms may be highly correlated), though it also discards some potentially useful information and can lead to undesirable results caused by spurious correlations.
  • Logistic Regression: A state-of-the-art supervised learning algorithm for machine learning that estimates the probability that a document is relevant, based on the features that it contains. In contrast to the Naïve Bayes, algorithm, Logistic Regression identifies features that discriminate between relevant and non-relevant documents.
  • Naïve Bayesian Classifier: A system that examines the probability that each word in a new document came from the word distribution derived from a trained responsive document or trained non-responsive documents. The system is naïve in the sense that it assumes that all words are independent of one another.
  • Neural Network: An Artificial Neural Network (ANN) is a computational model. It is based on the structure and functions of biological neural networks. It works like the way human brain processes information. It includes a large number of connected processing units that work together to process information.
  • Probabilistic Latent Semantic Analysis (PLSA): This is similar in spirit to LSA but it uses a probabilistic model to achieve results that are expected to be better.
  • Random Forests: An ensemble learning method for classification, regression, and other tasks, that operate by constructing a multitude of decision trees at training time and outputting the class that is the mode of the classes (classification) or mean prediction (regression) of the individual trees. Random decision forests correct for decision trees’ habit of overfitting to their training set.
  • Relevance Feedback: An active learning process in which the documents with the highest likelihood of relevance are coded by a human, and added to the training set.
  • Support Vector Machine: A mathematical approach that seeks to find a line that separates responsive from non-responsive documents so that, ideally, all of the responsive documents are on one side of the line and all of the non-responsive ones are on the other side.

General TAR Protocols (5,6,7,8)

Additionally, these technologies are generally employed as part of a TAR protocol which determines how the technologies are used. Examples of TAR protocols include:

  • Simple Passive Learning (SPL): In simple passive learning (“SPL”) methods, the teacher (i.e., human operator) selects the documents to be used as training examples; the learner is trained using these examples, and once sufficiently trained, is used to label every document in the collection as relevant or non-relevant. Generally, the documents labeled as relevant by the learner are re-reviewed manually. This manual review represents a small fraction of the collection, and hence a small fraction of the time and cost of an exhaustive manual review.
  • Simple Active Learning (SAL): In SAL methods, after the initial training set, the learner selects the documents to be reviewed and coded by the teacher, and used as training examples, and continues to select examples until it is sufficiently trained. Typically, the documents the learner chooses are those about which the learner is least certain, and therefore from which it will learn the most. Once sufficiently trained, the learner is then used to label every document in the collection. As with SPL, the documents labeled as relevant are generally re-reviewed manually.
  • Continuous Active Learning (CAL): In CAL, the TAR method developed, used, and advocated by Maura R. Grossman and Gordon V. Cormack, after the initial training set, the learner repeatedly selects the next-most-likely-to-be-relevant documents (that have not yet been considered) for review, coding, and training, and continues to do so until it can no longer find any more relevant documents. There is generally no second review because by the time the learner stops learning, all documents deemed relevant by the learner have already been identified and manually reviewed.
  • Scalable Continuous Active Learning (S-CAL): The essential difference between S-CAL and CAL is that for S-CAL, only a finite sample of documents from each successive batch is selected for labeling, and the process continues until the collection—or a large random sample of the collection—is exhausted. Together, the finite samples form a stratified sample of the document population, from which a statistical estimate of ρ may be derived.
  • Hybrid Multimodal Method: An approach developed by the e-Discovery Team (Ralph Losey) that includes all types of search methods, with primary reliance placed on predictive coding and the use of high-ranked documents for continuous active training.

Survey Background References

(1) Grossman, M. and Cormack, G. (2013). The Grossman-Cormack Glossary of Technology-Assisted Review. [ebook] Federal Courts Law Review. Available at: http://www.fclr.org/fclr/articles/html/2010/grossman.pdf [Accessed 31 Aug. 2018].

(2) Dimm, B. (2018). Expertise on Predictive Coding. [email].

(3) Roitblat, H. (2013). Introduction to Predictive Coding. [ebook] OrcaTec. Available at: https://theolp.wildapricot.org/Resources/Documents/Introduction%20to%20Predictive%20Coding%20-%20Herb%20Roitblat.pdf [Accessed 31 Aug. 2018].

(4) Tredennick, J. and Pickens, J. (2017). Deep Learning in E-Discovery: Moving Past the Hype. [online] Catalystsecure.com. Available at: https://catalystsecure.com/blog/2017/07/deep-learning-in-e-discovery-moving-past-the-hype/ [Accessed 31 Aug. 2018].

(5) Grossman, M. and Cormack, G. (2017). Technology-Assisted Review in Electronic Discovery. [ebook] Available at: https://judicialstudies.duke.edu/wp-content/uploads/2017/07/Panel-1_TECHNOLOGY-ASSISTED-REVIEW-IN-ELECTRONIC-DISCOVERY.pdf [Accessed 31 Aug. 2018].

(6) Grossman, M. and Cormack, G. (2016). Continuous Active Learning for TAR. [ebook] Practical Law. Available at: https://pdfs.semanticscholar.org/ed81/f3e1d35d459c95c7ef60b1ba0b3a202e4400.pdf [Accessed 31 Aug. 2018].

(7) Grossman, M. and Cormack, G. (2016). Scalability of Continuous Active Learning for Reliable High-Recall Text Classification. [ebook] Available at: https://plg.uwaterloo.ca/~gvcormac/scal/cormackgrossman16a.pdf [Accessed 3 Sep. 2018].

(8) Losey, R., Sullivan, J. and Reichenberger, T. (2015). e-Discovery Team at TREC 2015 Total Recall Track. [ebook] Available at: https://trec.nist.gov/pubs/trec24/papers/eDiscoveryTeam-TR.pdf [Accessed 1 Sep. 2018].

(9) Dimm, B. (2018). TAR, Proportionality, and Bad Algorithms (1-NN). [online] Clustify Blog – eDiscovery, Document Clustering, Predictive Coding, Information Retrieval, and Software Development. Available at: https://blog.cluster-text.com/2018/08/13/tar-proportionality-and-bad-algorithms-1-nn/ [Accessed 31 Aug. 2018].

(10) Robinson, R. (2013). Running Results: Predictive Coding One-Question Provider Implementation Survey. [online] ComplexDiscovery: eDiscovery Information. Available at: https://complexdiscovery.com/2013/03/05/running-results-predictive-coding-one-question-provider-implementation-survey/ [Accessed 31 Aug. 2018].

(11) Robinson, R. (2018). A Running List: Top 100+ eDiscovery Providers. [online] ComplexDiscovery: eDiscovery Information. Available at: https://complexdiscovery.com/2017/01/19/28252/ [Accessed 31 Aug. 2018].

(12) Robinson, R. (2018). Predictive Coding Technologies and Protocols: Overview and Survey [online] ComplexDiscovery: eDiscovery Information. Available at: https://complexdiscovery.com/predictive-coding-technologies-and-protocols-overview-and-survey/ [Accessed 16 Sep. 2018]

 

A Matter of Pricing? A Running Update of Semi-Annual eDiscovery Pricing Survey Responses

First administered in December of 2018 and conducted four times during the last two years with 334 individual responses, the semi-annual eDiscovery Pricing Survey highlights pricing on selected collection, processing, and review tasks. The aggregate results of all surveys as shared in the provided comparative charts may be helpful for understanding pricing and its impact on purchasing behavior on selected services over time.



Access the Results Now!

ComplexDiscovery combines original industry research with curated expert articles to create an informational resource that helps legal, business, and information technology professionals better understand the business and practice of data discovery and legal discovery.

All contributions are invested to support the development and distribution of ComplexDiscovery content. Contributors can make as many article contributions as they like, but will not be asked to register and pay until their contribution reaches $5.

A Competitive Advantage? FTC and DOJ Issue Antitrust Guidelines for Evaluating Vertical Mergers

According to FTC Chairman Joe Simons, “The new Guidelines reflect our...

Morae Acquires Legal Management Consultancy Janders Dean

According to Janders Dean founder Justin North, "Now more than ever,...

eDiscovery Mergers, Acquisitions, and Investments in Q2 2020

From UnitedLex to Onna, ComplexDiscovery findings, data points, and tracking information...

An Irish Update: DPC Ireland Publishes GDPR Regulatory Activity Report

The purpose of this two-year assessment is to provide a wider-angled...

A Running List: Top 100+ eDiscovery Providers

Based on a compilation of research from analyst firms and industry...

The eDisclosure Systems Buyers Guide – 2020 Edition (Andrew Haslam)

Authored by industry expert Andrew Haslam, the eDisclosure Buyers Guide continues...

The Race to the Starting Line? Recent Secure Remote Review Announcements

Not all secure remote review offerings are equal as the apparent...

Enabling Remote eDiscovery? A Snapshot of DaaS

Desktop as a Service (DaaS) providers are becoming important contributors to...

Home or Away? New eDiscovery Collection Market Sizing and Pricing Considerations

One of the key home (onsite) or away (remote) decisions that...

Revisions and Decisions? New Considerations for eDiscovery Secure Remote Reviews

One of the key revision and decision areas that business, legal,...

A Macro Look at Past and Projected eDiscovery Market Size from 2012 to 2024

From a macro look at past estimations of eDiscovery market size...

An eDiscovery Market Size Mashup: 2019-2024 Worldwide Software and Services Overview

While the Compound Annual Growth Rate (CAGR) for worldwide eDiscovery software...

Constrained Acceleration? The Summer 2020 eDiscovery Business Confidence Survey

Since January 2016, 2,089 individual responses to eighteen quarterly eDiscovery Business...

A Matter of Pricing? A Running Update of Semi-Annual eDiscovery Pricing Survey Responses

First administered in December of 2018 and conducted four times during...

A Pandemeconomic Indicator? Summer 2020 eDiscovery Pricing Survey Results

Based on the complexity of data and legal discovery, it is...

COVID-19 Constrained? The Impact of Six Issues on the Business of eDiscovery

In the spring of 2020, 51.2% of respondents viewed budgetary constraints...

Morae Acquires Legal Management Consultancy Janders Dean

According to Janders Dean founder Justin North, "Now more than ever,...

eDiscovery Mergers, Acquisitions, and Investments in Q2 2020

From UnitedLex to Onna, ComplexDiscovery findings, data points, and tracking information...

Mitratech Acquires CMPG Risk Solutions

According to the announcement, with the inclusion of enterprise and vendor...

Onna Raises $27M in Series B Funding

"Since we first invested in Onna last year, Slack deployed their...

Five Great Reads on eDiscovery for June 2020

From collection market size updates to cloud outsourcing guidelines, the June...

Five Great Reads on eDiscovery for May 2020

From review market sizing revisions to pandemeconomic pricing, the May 2020...

Five Great Reads on eDiscovery for April 2020

From business confidence to the boom of Zoom, the April 2020...

Five Great Reads on Data Discovery and Legal Discovery for March 2020

From business continuity considerations to cybersecurity attacks, the March 2020 edition...