FAQs

There may be multiple reasons for this. We found the following to be most likely:

• Your publication is too recent. Although we constantly update our database, we have to rely on the publisher of your article to inform the world about your contribution.
• Your publication has no DOI assigned to it. If it has, please email the DOI to us.
• Your publication is listed in a repository we are not aware yet. Send us a notice where we can find this repository.

There may be multiple reasons for this. We found the following to be most likely:

• The given summary is written like a review. Please try to be more specific describing the background, the methods, the central research question, the context and the underlying research topic.
• The provided abstract sounds like a political speech. Fair enough politicians talk the way they talk ... often without saying much. Our suggestion is to use more specific language in your input.
• Your input seems to vague. Do you recognise what the following text is about: "This paper is about new findings of recent research we conducted in our laboratory. Although latest methods have been used it was quite difficult to get to the results." You can do much better than that. Be more precise.

These are placeholder strings and values for information that is not initially provided by the publishers. You will most likely find the missing details on the original source.

We believe that context matters. Especially in science and scientific publications. In order to account for that we created a model to represent context of words, word groups, sentences, and paragraphs. Processing keywords without context would just generate nonsense results. Therefore we require you to provide at least 200 characters to add meaning to you input. Just imagine the word AIR. Depending on the context the word air is used in it may mean 1) a mixture of gasses, 2) the region above the ground, or 3) the medium for radio broadcast ... and probably even more. To avoid ambiguity and the typical time consuming iterative search with keywords, we search with context and provide you the relevant results on the first attempt.

"Poor English" is the language of science. Sorry for the joke. However, the overwhelming majority of publications is in English or contains English translations of title and/or abstract. That's why we decided to use English as our processing language. When you take a look at our changelog, we are planning to add a feature for auto-translation on non-english input. So we might get more flexible with other languages in the future.

Good point. We have included this as a feature from version 21.1.2 onwards. You may now specify up to 3 keywords or author names that must be included in the set of results. The keywords are applied using the logical "AND". Other Boolean operations are not yet implemented.

This happens every now and then. This has to do with the way publishers submit information. For some reason the publication appears under two different digital object identifiers (DOI) in the repositories. In other cases, a publication and its supplements are registerred under the same title but different DOI. This is out of our hands. It is not a duplicate in our database. When you see such a situation take a look at the publishers website to find out why this happens.

It seems you formulated your search input like a wikipedia article summary: Transmission (mechanics). Which is great when your aim was to explain something in "simple language". ODYSSAI compares you search input with other scientific publications and becomes most powerful when your search input reads like a scientific abstract. Please read through this wikipedia article to get a first impression on the structure of academic abstracts and reformulate your search input.

There isn't a strict answer to this. Most users said:

• 100% - 98% : Exact match. Such high values indicate that we have found the publication for which you provided us with the abstract.
• 97% - 90% : Superb match. In this range, publications fit very well. If it does not match the title for you, please look at the abstract.
• 89% - 80% : Good match. Here the results agree well and show a broader range of publications that are often used for writing reviews.
• 79% - 70% : Reasonable match. This range is rarely displayed in response to scientific abstracts. This can happen if keywords are specified that severely limit the selection.
• 69% - 50% : Match? We have hardly ever seen results in this range among the top 50. You may have entered a very rare combination of keywords in relation to your query.
• below 50% : This is like flipping a coin. We don't go there. ;)

Selecting and marking keywords enables the deep search. We still find matching publications to your query and filter them with the keywords provided. The keywords are treated with a logical "AND". This means they all must appear in the meta data of each result. Depending on the richness of the meta data this can be in title, abstract, author names or any combination of the three.

Yes, you can! Although we perform magic in the background don't expect too much since there are over 100 Million people called Wang on this planet. It is still more likely to get you the intended publications by describing the content you are looking for.

There are two options:

• Without keywords: If you have not specified must include keywords you triggered our interdisciplinary search. In that case you should get results above 50% similarity unless something went wrong on our side. Please send an email to support including your search query and a screen shot of your browser. We will immediately investigate that for you.
• With keywords: If you did specify must include keywords you triggered our deep search. In that case we filter the results to contain all provided keywords (logical "AND"). There may be indeed zero items that match your query AND contain all keywords. Please refine your search with different or less keywords.