Best Practices for Using Natural Language Search

NLPatent empowers users to explore patent documents using natural language queries, eliminating the need for laborious keyword-matching. By articulating an invention in natural language, our advanced Large Language Model (LLM) swiftly comprehends its scope, presenting tailored results with ease.


Despite its simplicity, there are some nuances in mastering the art of writing effective queries. Let's delve into these nuances together.

Writing a Great Natural Language Query 

Transitioning from keyword searching to natural language may require some adjustment. While listing specific keywords may feel intuitive, natural language search platforms interpret words based on the context within which they are used, so a more detailed description of the invention is necessary to provide meaning and interpretability by the AI.


Copy and pasting an invention disclosure or one or more claims provides an ample amount of detail for the AI to comprehend. However, the effectiveness of inputting claims directly will fluctuate based on the level of detail contained within them.


With any natural language query - whether it’s one that you’ve written up from scratch or repurposed from a client or patent document, it’s important to first evaluate its clarity and comprehensibility. Ask yourself whether someone with ordinary knowledge in the field could easily understand your description. Clear and concise queries tend to yield superior results, while convoluted or ambiguous language may diminish the quality of outcomes.


Here are some more key concepts to follow while writing a query.  

Use Full Sentences

Inputting full sentences is crucial for optimizing search results with our LLM. Unlike keyword-based approaches, our LLM relies on the context and meaning within a sentence to accurately interpret the claimed invention. Here are some examples:

Ineffective:

"Plant energy generation. Petal, photovoltaic cell."

Effective:

"A flower-shaped power plant energy generator. The device consists of a series of petals arranged around a central stem, mimicking the structure of a natural flower. Each petal is equipped with a photovoltaic cell that captures sunlight and converts it into electrical energy. As the petals open and close in response to environmental conditions, they generate a fluctuating electric current that can be stored in a battery for later use."

Provide More Context

Similar to the previous tip, our LLM best understands the scope of the search by being provided with more details/context. Try to be as descriptive as possible- imagine describing the invention to someone in real life who has some context about the field. 

Ineffective:

“An electric vehicle battery”

Effective:

“A Lithium ion battery that is used in electric vehicles, with a minimum charging capacity of 40 kWh. The battery is fire resistant and must have specific features in reducing its lifetime degradation over its lifetime.” 

Cut out Irrelevant Information

Sometimes an invention disclosure may contain too much information that is irrelevant to the parameters of the invention. If this is the case, it's best to ensure only core concepts are present in the query.

Ineffective:

The Multi-Purpose Pen (MPP) is a groundbreaking invention poised to redefine the way individuals approach writing tasks and beyond. With a plethora of features and capabilities the MPP transcends traditional writing instruments, offering parallel convenience and versatility. The MPP has a sleek aerodynamic silhouette, excluding sophistication and modernity making it a stylish accessory for professionals. Incorporating cutting-edge writing technology, it features state-of-the-art ballpoint pen mechanism engineered for fluid, skip-free ink delivery, guaranteeing a smooth writing experience with every stroke. Moreover its adaptive writing mechanism is engineered to dynamically adjust to the user's writing style and environment.  The MPP also contains a built-in digital voice recorder to enable users to capture audio notes, dictations and reminders on-the-go providing a convenient alternative to traditional pen-and-paper methods.

Effective:

The Multi-Purpose Pen (MPP) is a pen with several features including an adaptive writing mechanism engineered to dynamically adjust to the user's writing style and environment. Beyond the adaptive mechanism it also contains a built-in digital voice recorder to enable users to capture audio notes, dictations, and reminders on-the-go, which then can be accessed through an app. 

The later example contains only the inventive concepts whereas the ineffective query contains too much irrelevant information. 


While these 'ineffective' queries can still yield great results, incorporating these tips will enhance the quality of your search. Remember, these are guidelines, not rigid rules. Crafting effective queries is both an art and a science, so feel free to adjust and refine as you go. If you have the right concepts but need further refinement, consider leveraging the refinement feature rather than starting from scratch