In my previous article, I observed that while the numbers of Australian provisional applications filed in 2020 up until May had been down on 2019, filings in June and July were higher than during the same period last year, and that overall numbers have thus far shown greater resilience than in the last major economic downturn, i.e. the global financial crisis (GFC) of 2007-2009. I also noted back in May that many self-represented applicants, of both provisional and innovation patent applications, appeared to be directing their innovative attentions to problems arising out of the COVID-19 pandemic. I have been wondering, therefore, whether there may be a similar trend in provisional filing activity more generally that is helping to prop up the numbers, despite economic pressures associated with the pandemic.
I have now conducted some analysis, and it appears that there may be evidence to support this hypothesis. For the months of January to July, it appears that provisional filings associated with every industrial sector except chemistry are down compared with 2019 numbers, and that most of the growth within the chemistry sector can be attributed to pharmaceuticals. This is certainly consistent with an enhanced focus on healthcare, quite likely prompted by the global pandemic.
If you are familiar with the Australian patent system, you may already be wondering how I managed to analyse the sectors and fields of technology associated with provisional filings. Provisional applications are not published in full, and only limited bibliographic information is available, including the identity of the applicant, and the title of the application. Only a small fraction of these applications are filed by applicants whose industry sector and/or technology interests may be readily identified, most being filed by small private companies and individuals. That really leaves only the title as a means for ‘guessing’ the technology to which an application relates. Fortunately, thanks to machine learning technology, and the availability of a large amount of data relating to the classification of prior patent applications, we can do quite a bit better than just guessing!
On the assumption that many readers will not be as interested in the technical details of the machine learning approach, this article includes:
- a brief introduction to the machine learning model, sufficient to explain the classification system used, and generally how the model ‘learned’ to classify applications by title; and
- some results of analysing provisional filings over the past few years, which show a general decline in application numbers in most fields of technology, with the exceptions of those relating to ‘instruments’, which have been fairly flat, and ‘chemistry’, which has experienced a boost so far this year.
For those interested, in a separate article I provide additional technical detail of the machine learning model, and how it performed in testing and validation, as well as discussing how accurate we might expect it to be on the provisional application data.
Classifying Patent Applications by Technology Field and Sector
Before commencing any attempt at analysis, I needed to select a classification system – or a taxonomy, if you will – that can be consistently applied to all patent applications. The International Patent Classification (IPC) system is a hierarchical system of language independent symbols for the classification of patents by their associated areas of technology. The vast majority of patent applications filed in Australia are assigned one or more codes from the IPC by a (human) patent searcher or examiner, either in the Australian Patent Office or – in the case of international applications originally filed under the Patent Cooperation Treaty – in a patent office acting as the International Searching Authority. This is useful, because it provides a large body of data (around 500,000 applications, filed just in the relatively recent period since 2002) comprising application titles paired with IPC codes assigned by a human technology specialist.
The problem with IPC codes, however, is that they are far too specific at the lower levels of the hierarchy, while the structure at the higher levels of the hierarchy is not always well-aligned with industry sectors and fields of technology as we typically think about them in the worlds of commerce, research and development. Fortunately, the World Intellectual Property Organization (WIPO) has developed an ‘IPC concordance table’ that links the IPC code symbols with 35 fields of technology. The concordance table is updated (most recently in February 2016) to reflect revisions to the IPC. The 35 fields, which are divided into five broader industry sectors, are set out in the following table.
| No. | Sector | Field |
|---|---|---|
| 1 | Electrical engineering | Electrical machinery, apparatus, energy |
| 2 | Electrical engineering | Audio-visual technology |
| 3 | Electrical engineering | Telecommunications |
| 4 | Electrical engineering | Digital communication |
| 5 | Electrical engineering | Basic communication processes |
| 6 | Electrical engineering | Computer technology |
| 7 | Electrical engineering | IT methods for management |
| 8 | Electrical engineering | Semiconductors |
| 9 | Instruments | Optics |
| 10 | Instruments | Measurement |
| 11 | Instruments | Analysis of biological materials |
| 12 | Instruments | Control |
| 13 | Instruments | Medical technology |
| 14 | Chemistry | Organic fine chemistry |
| 15 | Chemistry | Biotechnology |
| 16 | Chemistry | Pharmaceuticals |
| 17 | Chemistry | Macromolecular chemistry, polymers |
| 18 | Chemistry | Food chemistry |
| 19 | Chemistry | Basic materials chemistry |
| 20 | Chemistry | Materials, metallurgy |
| 21 | Chemistry | Surface technology, coating |
| 22 | Chemistry | Micro-structural and nano-technology |
| 23 | Chemistry | Chemical engineering |
| 24 | Chemistry | Environmental technology |
| 25 | Mechanical engineering | Handling |
| 26 | Mechanical engineering | Machine tools |
| 27 | Mechanical engineering | Engines, pumps, turbines |
| 28 | Mechanical engineering | Textile and paper machines |
| 29 | Mechanical engineering | Other special machines |
| 30 | Mechanical engineering | Thermal processes and apparatus |
| 31 | Mechanical engineering | Mechanical elements |
| 32 | Mechanical engineering | Transport |
| 33 | Other fields | Furniture, games |
| 34 | Other fields | Other consumer goods |
| 35 | Other fields | Civil engineering |
I developed and trained a machine learning model using all Australian patent applications filed since 1 January 2002 having a valid title and primary IPC code assigned, by mapping the IPC code to one of the 35 fields of technology listed above. The resulting model, given a title as input, generates 35 output values, each of which represents the model’s estimate of the ‘probability’ that the title is associated with a corresponding one of the available fields of technology.
There are basically two things that can be done with this output. The first is to simply pick the largest output value, and ‘assign’ the title/application to the corresponding field of technology. This is called a ‘hard decision’ – not because it is difficult, but because it is uncompromising! Alternatively, all of the outputs can be retained, effectively ‘spreading’ the title/application across all of the fields of technology according to the ‘probabilities’ assigned by the model. This is called a ‘soft decision’, because it does not make a firm and final choice, but rather retains the additional information embodying the model’s uncertainty regarding the correct classification. In most cases, there is only a small number of fields with non-negligible output values, reflecting the reality that titles can be genuinely ambiguous, and also that many applications span multiple fields of technology.
By way of example, when I input the (fictional) title ‘a computer-implemented method for determining oxygen content of a blood sample’, the top four outputs are: analysis of biological materials (48%); computer technology (38%); medical technology (9.5%); and measurement (4.1%). I actually think that’s pretty amazing!
I used soft decisions in generating the results that follow. This means that each application makes a total contribution of ‘1’ to the outcomes (as it should), but that this contribution may be spread across multiple fields of technology.
Provisional Filings by Sector
The chart below shows the number of provisional filings, broken down by sector as determined by machine-classification of titles, over the months of January to July (i.e. the period for which 2020 data is available) during each year from 2017 to 2020. The results show a consistent decline in filings in the ‘electrical engineering’ and ‘mechanical engineering’ sectors, a fairly stable rate of filings for ‘instruments’, and a decline in ‘other’ fields that mostly took place between 2017 and 2018. The stand-out exception is ‘chemistry’ which, while relatively stable between 2017 and 2019, is the only sector to show growth in 2020.
Gains in the ‘chemistry’ sector are, however, insufficient to offset declines in the other sectors. As the following chart shows, overall provisional filing numbers (again, limited to the January-July period for fair comparison with 2020 figures) have been in steady decline since 2017.
Bucking the overall trend, provisional filings by self-represented applicants have actually increased slightly so far in 2020. If the past is anything to go by, most of these applications will lead nowhere. Nonetheless, in search of some additional insight into the activities of self-filers, the two charts below show the breakdown, by sector, of applications filed through agents (i.e. mostly patent attorneys) and those filed by self-represented applicants.
In general, self-represented applicants file relatively fewer applications relating to the ‘instruments’ and ‘chemistry’ sectors than those with professional representation. This is hardly surprising – most self-filers are individuals or micro-businesses who lack access to the facilities and resources to conduct serious R&D work in these areas. And yet there has been a spike in ‘chemistry’ applications, in particular, by self-filers in 2020. It is, however, likely that this spike merely reflects increased interest in ‘chemistry’ fields by self-filers, rather than any increase in capability. Examples of actual titles of self-filed applications having a top classification within the ‘chemistry’ sector include: ‘Disinfectant Aerosol Spray’; ‘Colour changing hand wash for the prevention of COVID-19 and other transmissible diseases’; ‘Article with Pathogen Inhibiting Treatment’; ‘An invention and method for enhanced screening’; and (just for something not pandemic-related) ‘Cannabinoid anti dandruff composition’.
Provisional Filings by Technical Field
The following chart shows the number of provisional filings, broken down into the 35 fields of technology as determined by machine-classification of titles, over the common period (i.e. the months of January to July) during each year from 2017 to 2020. Since it was not practical to include the full description of each field on this chart, I have used the field numbers, for which you may refer to the table above. The ‘chemistry’-related fields are numbered 14-24 and, as expected from the sector-level results, these include the only technical fields to show significant gains in 2020.
Drilling down into the ‘chemistry’ sector, the chart below shows the total number of provisional filings between January and July each year corresponding with the 11 technology fields within the sector. While there have been small gains in a number of these fields in comparison with 2019, the clear stand-out is ‘pharmaceuticals’, with growth of more than 50% over last year. And this does not appear to be a continuation of any pre-existing trend, since there was a decline in 2019 by comparison with filings in 2018.
The following chart shows classifications within the ‘chemistry’ sector for only those applications filed by self-represented applicants. It is clear that in most of these fields of technology self-filers make up a minority of all applicants. However, the significant spike in pharmaceutical filings is notable, with self-filers having contributed about half of all the additional applications filed in this field during the first seven months of 2020.
Conclusion – Is the Pandemic Offsetting its own Effect on Filings?
Based on the first seven months of filing data for 2020, the trend of declining numbers of provisional applications filed each year since 2017 seems set to continue. However, the year-to-date fall in filings would have been larger in 2020 were it not for the contribution of self-represented applicants, whose filing numbers are slightly up on 2019, and applicants who have developed a sudden interest in chemical – and, more particularly, pharmaceutical – technologies.
While we would typically expect an economic downturn to be accompanied by a downturn in new patent filings – which was certainly the case during the GFC – it seems that this effect is, for now at least, being partially offset by a desire to develop solutions to the non-economic problems presented by the COVID-19 pandemic. However, for the significant proportion of pandemic-inspired applications being filed by self-represented applicants, it is unlikely that substantial value will be generated in the longer term. And in the absence of the ‘pandemic effect’ contributing to perhaps a couple of hundred provisional applications, overall filing numbers so far this year might have been down by a further 5%-10% on 2019.
0 comments:
Post a Comment