ANNa-MD Demo: Multilingual Morphological Analogy Detection
Welcome to the ANNa-MD demo, your tool for exploring multilingual morphological analogy detection.
ANNa-MR Demo: Multilingual Morphological Analogy Resolution
Welcome to the ANNa-MR demo, your tool for exploring multilingual morphological analogy resolution.
Generation Model Demo: Multilingual Morphological Analogy Generation
Welcome to the Generation Model Demo, where you can explore the creation of multilingual morphological analogies.
- Select a Language: Choose your preferred language for the analogy analysis.
- Input Four Words: In the designated fields A, B, C, and D, enter four words in your selected language. If you're unsure or need an example, click the 'Prepare an example for me' button to automatically populate the fields with randomly selected words.
- Analyze the Analogy: Once you've filled in the four words, click the 'Is it an analogy?' button. The model will assess whether A : B :: C : D (with A, B, C, and D replaced by your inputs) forms a valid morphological analogy. For example, in English, 'cat : cats :: star : stars' is a valid morphological analogy because 'cat is to cats' as 'star is to stars'.
- Model Decision:The model will categorize the analogy as either 'Valid' or 'Invalid.'
- Model Confidence:You'll see the model's confidence in its decision, ranging from 0% (completely unsure) to 100% (certain).
How to Use
Output Information
The output will provide you with the following information:
Explore Model Statistics
For an even deeper understanding of the model's performance, you can view statistics on different model versions by clicking the 'Model Statistics' button.
Please Note
Our platform is a work in progress, and we are continually working to enhance the models' performance. You may experience a processing time of approximately 15 to 17 seconds to obtain results from our classification models. Thank you for your patience as we improve and refine our service.
- Select a Language: Choose the language in which you want to explore morphological analogies.
- Input Three Words: In the designated fields A, B, and C, enter three words in your selected language. If you're uncertain or need an example, click the 'Prepare an example for me' button to auto-fill the fields with randomly selected words
- Analyze the Analogy: Once you've filled in the three words and chosen your distance metric, click the 'Get closest solution' button. The model will attempt to complete A : B :: C : ? (with A, B, and C replaced by your inputs) by suggesting a word D. The goal is to form a valid analogy where A : B :: C : D. For example, in English, 'cat : cats :: star : ?' would output 'stars' because 'cat is to cats' as 'star is to stars.'
- Closest Known Word: The model's suggested word in place of D.
- Distance to the Word: The distance in the embedding space between the model's output and the embedding of the selected solution. The system also displays the five most similar results with their cosine similarity.
- Result Message: You'll receive a message describing the result or an error message if necessary
How to Use
Output Information
The output will provide you with the following information:
Website Status
Please be aware that our website is a work in progress, and some demo functionalities may not be available yet.
Note
These retrieval models may exhibit biases and may occasionally produce erroneous results. We are continuously working to improve their performance and accuracy. Thank you for your understanding as we strive to enhance this service. Explore ANNa-MR and uncover morphological analogies in your chosen language!
- Select a Language: Choose the language in which you want to explore morphological analogies.
- Input Three Words: In the designated fields A, B, and C, enter three words in your selected language. If you're unsure or need an example, click the 'Prepare an example for me' button to have the fields automatically filled with randomly selected words.
- Analyze the Analogy: After you've provided the three words, click the 'Generate' button. The model will endeavor to complete A : B :: C : ? (with A, B, and C replaced by your inputs) by generating a word D. The objective is to establish a valid analogy, such as in English, where 'cat : cats :: star : ?' would yield 'stars' because 'cat is to cats' as 'star is to stars.'
- Generated Word: The word generated by the model to complete the analogy, occupying the position of D.
- Similar Words: A set of the five most similar words generated by the model, along with their respective confidence levels expressed as percentages.
- Computation Time: Please be aware that the model requires approximately 15 to 20 seconds to produce a result.
How to Use
Output Information
The output will provide you with the following information:
Website Status
Our website is continuously evolving, and we may be adding new features and improvements. Thank you for your patience as we work to enhance our service.
Unlock the power of analogical thinking with the Generation Model Demo and create morphological analogies in your chosen language!
Note: Our models may not always provide precise results, as they can be inconsistent. We're actively working on improvements and will keep you updated on our progress.
Computing in progress ...
The closest generated result for:
: :: →Five most similar answers:
The closest result found for:
Paris : France :: Berlin → GermanyFive most similar answers:
The generated analogy for:
Paris : France :: Berlin : GermanyValid
(0%)