Patent paperwork usually use authorized and extremely technical language, with context-dependent phrases that will have meanings fairly totally different from colloquial utilization and even between totally different paperwork. The method of utilizing conventional patent search strategies (e.g., key phrase looking out) to go looking by the corpus of over 100 million patent paperwork may be tedious and lead to many missed outcomes because of the broad and non-standard language used. For instance, a “soccer ball” could also be described as a “spherical recreation gadget”, “inflatable sportsball” or “ball for ball recreation”. Moreover, the language utilized in some patent paperwork might obfuscate phrases to their benefit, so extra highly effective pure language processing (NLP) and semantic similarity understanding may give everybody entry to do an intensive search.
The patent area (and extra normal technical literature like scientific publications) poses distinctive challenges for NLP modeling because of its use of authorized and technical phrases. Whereas there are a number of generally used general-purpose semantic textual similarity (STS) benchmark datasets (e.g., STS-B, SICK, MRPC, PIT), to the very best of our data, there are at the moment no datasets targeted on technical ideas present in patents and scientific publications (the considerably associated BioASQ problem accommodates a biomedical query answering process). Furthermore, with the persevering with development in measurement of the patent corpus (tens of millions of recent patents are issued worldwide yearly), there’s a have to develop extra helpful NLP fashions for this area.
Right now, we announce the discharge of the Patent Phrase Similarity dataset, a brand new human-rated contextual phrase-to-phrase semantic matching dataset, and the accompanying paper, offered on the SIGIR PatentSemTech Workshop, which focuses on technical phrases from patents. The Patent Phrase Similarity dataset accommodates ~50,000 rated phrase pairs, every with a Cooperative Patent Classification (CPC) class as context. Along with similarity scores which can be usually included in different benchmark datasets, we embrace granular ranking courses just like WordNet, corresponding to synonym, antonym, hypernym, hyponym, holonym, meronym, and area associated. This dataset (distributed underneath the Artistic Commons Attribution 4.0 Worldwide license) was utilized by Kaggle and USPTO because the benchmark dataset within the U.S. Patent Phrase to Phrase Matching competitors to attract extra consideration to the efficiency of machine studying fashions on technical textual content. Preliminary outcomes present that fashions fine-tuned on this new dataset carry out considerably higher than normal pre-trained fashions with out fine-tuning.
The Patent Phrase Similarity Dataset
To raised practice the following era of state-of-the-art fashions, we created the Patent Phrase Similarity dataset, which incorporates many examples to handle the next issues: (1) phrase disambiguation, (2) adversarial key phrase matching, and (3) arduous destructive key phrases (i.e., key phrases which can be unrelated however obtained a excessive rating for similarity from different fashions ). Some key phrases and phrases can have a number of meanings (e.g., the phrase “mouse” might discuss with an animal or a pc enter gadget), so we disambiguate the phrases by together with CPC courses with every pair of phrases. Additionally, many NLP fashions (e.g., bag of phrases fashions) won’t do properly on information with phrases which have matching key phrases however are in any other case unrelated (adversarial key phrases, e.g., “container part” → “kitchen container”, “offset desk” → “desk fan”). The Patent Phrase Similarity dataset is designed to incorporate many examples of matching key phrases which can be unrelated by adversarial key phrase match, enabling NLP fashions to enhance their efficiency.
Every entry within the Patent Phrase Similarity dataset accommodates two phrases, an anchor and goal, a context CPC class, a ranking class, and a similarity rating. The dataset accommodates 48,548 entries with 973 distinctive anchors, cut up into coaching (75%), validation (5%), and check (20%) units. When splitting the info, all the entries with the identical anchor are stored collectively in the identical set. There are 106 totally different context CPC courses and all of them are represented within the coaching set.
|acid absorption||absorption of acid||B08||actual||1.0|
|acid absorption||acid immersion||B08||synonym||0.75|
|acid absorption||chemically soaked||B08||area associated||0.25|
|acid absorption||acid reflux disease||B08||not associated||0.0|
|gasoline mix||petrol mix||C10||synonym||0.75|
|gasoline mix||gas mix||C10||hypernym||0.5|
|gasoline mix||fruit mix||C10||not associated||0.0|
|faucet meeting||water faucet||A22||hyponym||0.5|
|faucet meeting||water provide||A22||holonym||0.25|
|faucet meeting||college meeting||A22||not associated||0.0|
|A small pattern of the dataset with anchor and goal phrases, context CPC class (B08: Cleansing, C10: Petroleum, gasoline, gas, lubricants, A22: Butchering, processing meat/poultry/fish), a ranking class, and a similarity rating.|
Producing the Dataset
To generate the Patent Phrase Similarity information, we first course of the ~140 million patent paperwork within the Google Patent’s corpus and robotically extract necessary English phrases, that are usually noun phrases (e.g., “fastener”, “lifting meeting”) and purposeful phrases (e.g., “meals processing”, “ink printing”). Subsequent, we filter and maintain phrases that seem in not less than 100 patents and randomly pattern round 1,000 of those filtered phrases, which we name anchor phrases. For every anchor phrase, we discover all the matching patents and all the CPC courses for these patents. We then randomly pattern as much as 4 matching CPC courses, which grow to be the context CPC courses for the particular anchor phrase.
We use two totally different strategies for pre-generating goal phrases: (1) partial matching and (2) a masked language mannequin (MLM). For partial matching, we randomly choose phrases from your entire corpus that partially match with the anchor phrase (e.g., “abatement” → “noise abatement”, “materials formation” → “formation materials”). For MLM, we choose sentences from the patents that include a given anchor phrase, masks them out, and use the Patent-BERT mannequin to foretell candidates for the masked portion of the textual content. Then, all the phrases are cleaned up, which incorporates lowercasing and the elimination of punctuation and sure stopwords (e.g., “and”, “or”, “mentioned”), and despatched to knowledgeable raters for overview. Every phrase pair is rated independently by two raters expert within the know-how space. Every rater additionally generates new goal phrases with totally different scores. Particularly, they’re requested to generate some low-similarity and unrelated targets that partially match with the unique anchor and/or some high-similarity targets. Lastly, the raters meet to debate their scores and give you remaining scores.
To guage its efficiency, the Patent Phrase Similarity dataset was used within the U.S. Patent Phrase to Phrase Matching Kaggle competitors. The competitors was extremely popular, drawing about 2,000 rivals from all over the world. A wide range of approaches have been efficiently utilized by the highest scoring groups, together with ensemble fashions of BERT variants and prompting (see the total dialogue for extra particulars). The desk beneath reveals the very best outcomes from the competitors, in addition to a number of off-the-shelf baselines from our paper. The Pearson correlation metric was used to measure the linear correlation between the anticipated and true scores, which is a useful metric to focus on for downstream fashions to allow them to distinguish between totally different similarity scores.
The baselines within the paper may be thought of zero-shot within the sense that they use off-the-shelf fashions with none additional fine-tuning on the brand new dataset (we use these fashions to embed the anchor and goal phrases individually and compute the cosine similarity between them). The Kaggle competitors outcomes display that by utilizing our coaching information, one can obtain important enhancements in contrast with present NLP fashions. Now we have additionally estimated human efficiency on this process by evaluating a single rater’s scores to the mixed rating of each raters. The outcomes point out that this isn’t a very simple process, even for human specialists.
|Kaggle 1st place single||High quality-tuned||0.87|
|Kaggle 1st place ensemble||High quality-tuned||0.88|
|Efficiency of standard fashions with no fine-tuning (zero-shot), fashions fine-tuned on the Patent Phrase Similarity dataset as a part of the Kaggle competitors, and single human efficiency.|
Conclusion and Future Work
We current the Patent Phrase Similarity dataset, which was used because the benchmark dataset within the U.S. Patent Phrase to Phrase Matching competitors, and display that by utilizing our coaching information, one can obtain important enhancements in contrast with present NLP fashions.
Extra difficult machine studying benchmarks may be generated from the patent corpus, and patent information has made its approach into lots of at present’s most-studied fashions. For instance, the C4 textual content dataset used to coach T5 accommodates many patent paperwork. The BigBird and LongT5 fashions additionally use patents through the BIGPATENT dataset. The supply, breadth and open utilization phrases of full textual content information (see Google Patents Public Datasets) makes patents a singular useful resource for the analysis group. Potentialities for future duties embrace massively multi-label classification, summarization, data retrieval, image-text similarity, quotation graph prediction, and translation. See the paper for extra particulars.
This work was doable by a collaboration with Kaggle, Satsyil Corp., USPTO, and MaxVal. Due to contributors Ian Wetherbee from Google, Will Cukierski and Maggie Demkin from Kaggle. Due to Jerry Ma, Scott Beliveau, and Jamie Holcombe from USPTO and Suja Chittamahalingam from MaxVal for his or her contributions.