Resources / Blog

Patent Embodiment: Why Multiple Are So Helpful

Jul 25, 2018

Patent embodiment is the description of the production, use, practice, or expression of an invention in the patent application.” according to the UpCounsel. Imagine this: ‘Company A’ owns a patent for an energy-efficient lawnmower. ‘Company B’ starts manufacturing a lawnmower that looks eerily similar. When ‘Company A’ finds out, it sends a cease and desist letter. ‘Company B’ fights back, saying its product is not infringing. They go to court, where a deciding authority — a judge, examiner or board — delves into ‘Company A’s’ patent claims to determine what legally constitutes infringing activity.

The claims’ language is at the core of this decision: If it is solid and transparent, then the semantic investigation will probably work out in favor of ‘Company A,’ the patent owner. If the patent claims are vague, ‘Company B’ could win the case.

How to protect yourself

To avoid being on the losing end of such a judgment, IP applicants should consider listing multiple embodiments on their patent applications. The Oxford Dictionary defines an embodiment as “representing or expressing something in a tangible or visible form.” On a patent application, an embodiment is a detailed description of how an invention can be made or used. It’s not enough to assume that this is obvious or implied – it must be explicitly stated to take advantage of maximum protection.

For example, if there are multiple methods of making an invention or if it is an invention that has (or could have) various uses, the patent applicant should list every possible embodiment they can imagine in the specification. Doing so makes it less likely that an infringer will later be able to exploit vagueness in the issued patent.

Global Patent Filing

Remember Companies A and B above? Now that you understand how critical patent semantics can be, one of the most important steps you can take when filing for protection in multiple countries is to ensure that you have translated every embodiment accurately. You want there to be as little interpretive “wiggle room” as possible should IP litigation occur.

IP litigation can be particularly challenging when a determination is to be made by a judge, examiner or board that is familiar with a different IP system, language and culture from your own. An LSP specializing in international patent filing can ensure that every application you file is translated accurately and appropriately.

Case Study: Purdue vs. Alvogen

In the case Purdue Pharma LP vs. Alvogen Pine Brook LLC, one point of contention clearly illustrates the importance of thoroughly listing multiple embodiments. Here’s what happened:

  • Purdue Pharma owned US Patent No. 9,750,703 for “encased tamper-resistant controlled release dosage forms.”
  • Purdue Pharma used this technology to create Butrans, a prescription medication used to “treat pain that continues around the clock.” Butrans comes in the form of a patch that contains an extended-release opioid absorbed through the skin over seven days.
  • While Butrans was still patent-protected, Alvogen, another life science company, started marketing a generic version of the extended-release pain medication.
  • Purdue Pharma claimed Alvogen’s generic version infringed upon Patent No. 9,750,703.
  • The case went to the Delaware District Court, where Purdue won.

A closer look at Patent No. 9,750,703

In discussing a time-release shell for a pharmaceutical, the following phrase in Patent No. 9,750,703 became critical to Purdue’s win: “A layer encasing the core and comprising a second portion of hydrocodone bitartrate dispersed in a second matrix material.” At issue was the meaning of “layer encasing the core.” Did the patent refer to, and protect only a single-material layer, as Alvogen asserted? If so, it couldn’t apply to a shell made up of multiple materials, as their product did.

Ultimately, the court concluded after reviewing the following embodiment regarding different ways Purdue’s shell could be constructed: “The shell of the dosage form can be formed, e.g., by compression coating, molding, spraying one or more layers onto the core, dipping one or more layers onto the core or a combination thereof.” The court sided with Purdue and concluded that the term ‘layer encasing the core’ means “one or more materials enclosing a space or surrounding the core.”

The takeaway here is that because Purdue’s application was farsighted enough to include these multiple embodiments — i.e., the various ways in which their shell could be created — they were able to prevail in the infringement suit against Alvogen.

Invaluable patent embodiment: How multiple patent embodiments are especially crucial with global patents

If you plan to protect your patents in multiple countries, the last place you want to find yourself is in a legal dispute in front of a foreign authority who has to read between your lines. Working with an LSP that can anticipate potential local unclarities can, ultimately, make a huge difference. By knowing each region’s IP system, being fluent in the local language, and understanding the cultural lens through which the details in your filing may one day be viewed, the right LSP can help ensure your IP protection is rock-solid uncontestable.

Resources / Blog

Understanding the Impact of AI and Big Data on the Language Industry

Jul 18, 2018

In today’s data-driven world, the technology sector is abuzz with the potential of Artificial Intelligence (AI) and the power of Big Data. It’s exciting, futuristic stuff, and offers all sorts of new opportunities for a variety of services, including language service providers (LSPs).

How big is Big Data?

When we talk about Big Data, we mean big data — 2.5 quintillion bytes of data are added to the world’s digital repositories every day. In 2015, roughly 7.9 zettabytes were collected (1 zettabyte = 1 trillion gigabytes), and we’re expected to reach 176 zettabytes by 2025.

All of this data creates huge sets of structured and unstructured digital information that comes from pretty much everywhere. For example, big data sets can include your past sales data, online buyer reviews, real-time streaming, scientific experiments, government censuses, and pretty much anything else you can think of.

What do we get from AI and Big Data?

At its purest level, artificial intelligence is the product of thinking that is done by computer software. For now, practical use of AI is found chiefly in programs that utilize machine learning and deep learning – both of which use algorithms to parse data and then make intelligent decisions (and even build new neural networks) based on what was learned.

Big Data is about analysis, or as The ATA Chronicle puts it, “taking large amounts of data and using software tools to identify previously undiscovered patterns, trends, correlations, and associations.”

For LSPs, one benefit of applying AI learning to Big Data could be early identification of changing industry trends – which could result in better advanced planning for customers’ changing needs.

Consumer benefits

Some experts believe AI’s benefits will result in more people being able to take advantage of translation and localization services as they become more affordable. The hope, ultimately, is that more companies, organizations, and individuals will be able to converse with others throughout the world.

The historical model these experts cite is the clothing industry, where the upheaval of industrialization eventually revolutionized the industry by increasing the speed of production and lowering its cost, enabling an explosive increase in the sales of lower-cost goods.

Are we there yet?

The short answer is “no,” though some, such as VentureBeat, predict AI will mature pretty quickly and be a formidable force in translations by 2024. For now, though, human intervention in translations, especially regulated translations, is absolutely mandatory.

In fact, a man vs. machine competition demonstrated just that in South Korea last year. In this competition, four human translators were each given a different piece of content (under 250 words) to translate. They each had 50 minutes to complete the work while three AI translation programs (Naver’s Papago, Google Translate and Systran) were simultaneously assigned the same four jobs. Competition scoring was based on accuracy, language expression, logic and organization. Despite the fact that each one of the AI programs beat the human translators in turnaround time, it was still a worst-case scenario for AI.

“Some 90 percent of the artificial intelligence programs’ translations had grammatical errors,” said Kwak Jung-chul, president of the Korean Association of Translators and Interpreters. “The programs failed to understand the message that writers were trying to convey.”

The future of AI in language services

Experts have been weighing in on what we can expect as AI use grows in modern business. One especially interesting area to look at is web robots (also called chatbots or just bots). You’ve probably come across one of these while shopping on Amazon, surfing Facebook or working inside a Microsoft platform – and chances are you will only see more of them as time goes on.

Customer service chatbots are, according to Forbes, “[an] evolving technology [that] makes it easier for businesses to engage their customers, cut down on their workload and proactively prevent minor customer service issues from becoming major problems.” The value of a customer service bot, however, depends entirely on the quality of the automated script on which the bot runs. High quality scripts in multiple languages can result in higher customer satisfaction as well as the opportunity to obtain new business from different markets.

Push to pull translation

Andrew Joscelyne, writing for the Globalization and Localization Association (GALA), points out another interesting trend that could have a major influence on the future of the translation industry. He calls the shift “push to pull translation.”

Here’s what that means: Traditionally, companies have doled out (pushed) information that they consider to be important – and only translated it into the languages they’ve deemed profitable. This is the “push” side of the equation – the company chooses the type of information to distribute and the language/s to distribute it in. However, as online audiences grow, they may want to “pull” information (including languages) that companies don’t yet offer.

For example, a growing company might receive requests for business in 30 different languages – but it only has translators for 3 of those languages. That’s a major loss of potential business! Big Data can help companies look at users’ digital profiles, geolocations and personalized habits to help determine which additional languages and information could increase business the most.

Challenges and opportunities

It’s clear, then, that AI and Big Data stand to transform language services and are already beginning to do so. Once LSPs learn to leverage the wisdom buried in Big Data and begin to take advantage of more advanced AI, all of us will be able to communicate more freely with other people in other places.

For now, however, human translators continue to be a necessity in the fields of business, law, life sciences and intellectual property. In these regulated industries, quality simply can’t be sacrificed for convenience.