by Bernardo Bravo, Senior Data Scientist, Yewno,
and Ali Limon, Senior Data Scientist, Yewno


In this white paper we present how investors can leverage the Yewno Knowledge Graph Concept Exposure data to create thematic investment strategies based on concept exposures related to specific themes.  In other words, a concept-based thematic strategy. For this purpose, we leverage the patents layer of the Graph and its derived scores provided in our Alternative Data API.

The strategies select securities from the Nasdaq US Benchmark Index (NQUSB) constituents, based on the Aggregate Exposure Score to a set of concepts related to Future Mobility. We present a long only monthly-rebalanced strategy that outperforms several selected Future Mobility benchmarks on a risk adjusted basis during the time period considered, by as much as 6.2% and 5.5% per annum on net return compared to KARS and DRIV ETFs, accordingly.

Introduction and Rationale

The Yewno Patent Network layer is a new, cutting edge alternative data product suite that provides unparalleled insight into company intellectual property innovation. Leveraging our patented Yewno Knowledge Graph and AI-driven entity-linking technology, Patent Analytics provides point-in-time entity-specific exposure scores to key technological and environmental themes. Patents are a uniquely forward-looking data source which allows our users to identify companies developing and investing in new disruptive technologies. These include thousands of concepts including  Electric Cars, Blockchain, Alternative Energy and Robotics,. The Yewno Patents data sources include two jurisdictions: published patents from the WIPO (International Patent Office) and the USPTO (US Patent Office)

The Transportation sector has been widely impacted by innovation and technology over the last decade. New developments made possible the creation and implementation of frameworks to drive the production of autonomous vehicles, electric vehicles and robots.  In today’s highly intelligent automation environment, companies that are leveraging these technologies are generating considerable economic and shareholder value. These companies are able to sustain growth, wide margins, and by extension a competitive advantage over their peers in the transportation industry. 

In this use case we leverage the Yewno Patents Scores alternative data feed to construct a factor representing the degree of exposure of companies to the theme Future Mobility.  Specifically, we define a set of Future Mobility related concepts which defines the theme and use the average Aggregate score to construct a portfolio of companies exposed to it.  The Aggregate Concept Exposure score which is a linear combination of the other Concept Exposure scores: contribution, pureplay, centrality and similarity [1]

Data Overview

Concept Exposures API query

The data used in this backtesting exercise is concept exposure scores generated from patent sources from 2018-05 to 2021-06 for all companies in the NQUSB Index. The lookback window used as a parameter is 180d since we are looking to test factors in a monthly frequency. Nine concepts are used for this strategy:

72df109d5f209be31fe63a3a693a129b ———– Charging Station

b9f5a77a153d927c5eb48dfaa1cab044 ———- Electric Car

930c3e7d254ba2fbeb2e042af59e5c1f  ———- Autonomous Car

7cd45b9c3aba5de164c8faa5dca7d77d ———- Robot

d178b1cb541f10d2c4b15436681801a5 ——— Chipset

f1bea51ed2b5866accd0c2b1a03e6857 ——— Smart Mobility Architecture

b8393b891675b38d06603577857fcd36 ——— Battery (electricity)

c3efde11f903b8609b115ec804c30f4d   ———-  Lithium-ion battery

API call example 


“sourceType”: “patents”,

“window”: “180d”,

“dateStart”: “2016-10-01”,

“dateEnd”: “2021-05-30”,

   “source”: {

       “concepts”: [


          “930c3e7d254ba2fbeb2e042af59e5c1f”, “7cd45b9c3aba5de164c8faa5dca7d77d”,

          “d178b1cb541f10d2c4b15436681801a5”, “f1bea51ed2b5866accd0c2b1a03e6857”,

          “57bc6d36f8d882ea8cbf75803fbff329”, “b8393b891675b38d06603577857fcd36”,


          “0a2e23a3050e7dd612332c60dee7948a”, “6df2ad17b8bbc6ba8a0683ff3bd5df53”,




   “expandIsins”: true,

   “expandParents”: true,

   “filters”: {

       “entitiesOnly”: true



Thematic Strategies

Long only strategies

We queried the concept exposures API to retrieve all historical data by concept and company. For each data point, we performed a transformation to aggregate factors for all concepts by date and company using the average of Aggregate Score.

We create the portfolio using the corresponding Aggregate Score tilted by market cap. The cumulative returns are shown in Figure 1. The strategy outperforms two selected benchmarks, KARS and DRIV, by as much as 6.2% and 5.5% per annum on net return, respectively,during backtesting period.

Cumulative returns on long only strategies for Future Mobility Strategy and selected Benchmarks

Cumulative returns on long only strategies for Future Mobility Strategy and selected Benchmarks

Sources: Factset, Morningstar, Yewno, Yahoo Finance

Figure 1: Cumulative returns on long only strategies for Future Mobility Strategy and selected Benchmarks


The Aggregated Patents Exposure Score reflects the extent to which companies are either directly or indirectly connected to Future Mobility concepts in the Knowledge Graph induced by patents mentions. In particular, a higher aggregate score reflects companies that are publishing more patents in the related technologies and are connected to companies or technologies with high relevance in the Future Mobility space.

In Figure 2 we present the performance comparison of the Future Mobility strategy and selected benchmarks. Over the whole period the strategy outperformed the benchmarks and it had lower volatility yielding a higher sharpe ratio. The strategy outperforms the benchmarks by risk/return ratio of 0.4 and 0.35 compared to KARS and DRIV, respectively. 

Performance metrics for Future Mobility Strategy and Benchmarks

Figure 2:Performance metrics for Future Mobility Strategy and Benchmarks

We also show the top average holdings in the last year, observing that some companies such as Apple, Tesla, and Intel published patents in the previous year explicitly forming part of some Future Mobility Innovation. 

Top avg Holdings of Future Mobility Strategy in the last yearFigure 3: Top avg Holdings of Future Mobility Strategy in the last year


Explaining General Motors Exposure

Aggregate Exposure is computed as a linear combination of five scores derived from the Patents Graph, aggregate score was designed to weight higher first order connections via Pure Play and Contribution scores. Companies with higher aggregate scores in the patent layer are expected to have significantly higher numbers of published patents than those with low scores. In particular it was detected by the Knowledge Graph that General Motors publish patents related to different fields via subsidiaries such as GM Cruise Holdings and GM Global Technology, among others. Below we include an example explaining the connection between Charging Station and GM Global Technology.  We observe that the document presents the development of mobile charging stations for electric-drive vehicles, which are technologies relevant for Future Mobility strategy. 


Patent Link 


Explaining Johnson & Johnson Exposure

Similar to GM, Johnson and Johnson has shown significant activity in patent publications via different subsidiaries. In particular Ethicon is a company that concentrates its R&D in technologies such as Electric Motors, Charging Station, Chips, Batteries and Robots for surgical devices.  Even if most of Ethicon inventions are applicable to the Healthcare industry, these selected technologies also have a relevant relationship with Electrification in the Transportation industry.

PATENT 2Patent Link


In this example we constructed a trading strategy using Knowledge Graph Exposure Scores on patents, providing an indication that the Yewno Knowledge Graph alternative data exhibits signals that can be used systematically for thematic investment.  The long strategy constructed for Future Mobility using patent data had superior risk adjusted performance to the selected benchmarks, achieving higher returns and lower volatility in the tested period. 

Patents are a unique data source that provide forward looking and unbiased information of companies’ innovation portfolios. In today’s technology-driven world, companies allocate significant resources to developing technologies such as autonomous vehicles, drones, artificial intelligence, big data and alternative energy. Intellectual property derived signals represent huge opportunities for portfolio investing because they can highlight companies with the potential to disrupt industries. Patents data can be used as standalone signals and as a source to augment existing strategies for portfolio construction.  



Leave a Reply