Information synthesis engine

One or more machine-readable storage media containing instructions for use in generating a report from multiple sources, the one or more machine-readable storage media being one or more tangible media, the instructions being executable by one or more processing devices to:

• obtain a topic by processing input queries relating to the topic, the input queries corresponding to queries from multiple users made over a predefined period of time;
• obtain information about the topic from the multiple sources, the information comprising excerpts from the multiple sources that meet one or more criteria;
• wherein obtaining the information comprises:
• if the topic is ambiguous, disambiguating the topic by combining names of the one or more categories with the topic, thereby producing a disambiguated topic and formulating a search query based on the disambiguated topic;
• if the topic is not ambiguous, formulating the search query based on the topic;
• and
• searching the multiple sources using the search query to obtain the information;
• and
• generate the report using the excerpts, wherein generating the report comprises:
• obtaining subtopics by analyzing the excerpts from the multiple sources for commonalities;
• organizing the excerpts and hyperlinks associated with the excerpts under appropriate subtopics to provide a narrative flow relating to the topic, the excerpts being organized by subtopic and, in a subtopic, being organized in an order;
• and
• editing text in the excerpts, wherein editing comprises editing text in excerpts so that excerpts from different sources under subtopics read more like the excerpts from different sources came from a single source;
• and
• display the report.

A method, performed on one or more processing devices, for generating a report from multiple sources, the method comprising:

• (I) obtaining a topic based on input queries relating to the topic;
• (II) obtaining information about the topic from the multiple sources, the information comprising excerpts from the multiple sources that meet one or more criteria;
• and
• (III) generating the report using the excerpts, wherein generating the report comprises:
• (i) obtaining subtopics for the excerpts;
• (ii) organizing the excerpts based on the subtopics;
• and
• (iii) editing text in the excerpts;
• (IV) wherein obtaining information about the topic comprises:
• (i) assigning the topic to one or more categories;
• (ii) determining whether the topic is ambiguous based on the one or more categories to which the topic is assigned;
• (iii) wherein, if the topic is determined to be ambiguous, the method comprises:
• disambiguating the topic by combining names of the one or more categories with the topic, thereby producing a disambiguated topic;
• and
• formulating a search query based on the disambiguated topic;
• (iv) wherein, if the topic is determined not to be ambiguous, the method comprises:
• formulating the search query based on the topic;
• (v) searching the multiple sources using the search query;
• (vi) receiving search results as a result of searching the multiple sources, the search results comprising documents corresponding to the search query;
• (vii) selecting a subset of the documents from which to extract the excerpts, the subset of the documents being selected based on intrinsic metrics and extrinsic metrics, the intrinsic metrics comprising data from the documents and the extrinsic metrics comprising data relating to the documents but not from the documents;
• and
• (viii) extracting the excerpts from the subset the documents, wherein the excerpts are extracted from the subset of documents based on the one or more criteria, the excerpts comprising text and non-text.

One or more machine-readable storage media containing instructions for use in generating a report from multiple sources, the one or more machine-readable storage media being one or more tangible media, the instructions being executable by one or more processing devices to:

• (I) obtain a topic based on input queries relating to the topic;
• (II) obtain information about the topic from the multiple sources, the information comprising excerpts from the multiple sources that meet one or more criteria;
• and
• (III) generate the report using the excerpts, wherein generating the report comprises:
• (i) obtaining subtopics for the excerpts;
• (ii) organizing the excerpts based on the subtopics;
• and
• (iii) editing text in the excerpts;
• (IV) wherein obtaining information about the topic comprises:
• (i) assigning the topic to one or more categories;
• (ii) determining whether the topic is ambiguous based on the one or more categories to which the topic is assigned;
• (iii) wherein, if the topic is determined to be ambiguous, the method comprises:
• disambiguating the topic by combining names of the one or more categories with the topic, thereby producing a disambiguated topic;
• and
• formulating a search query based on the disambiguated topic;
• (iv) wherein, if the topic is determined not to be ambiguous, the method comprises:
• formulating the search query based on the topic;
• (v) searching the multiple sources using the search query;
• (vi) receiving search results as a result of searching the multiple sources, the search results comprising documents corresponding to the search query;
• (vii) selecting a subset of the documents from which to extract the excerpts, the subset of the documents being selected based on intrinsic metrics and extrinsic metrics, the intrinsic metrics comprising data from the documents and the extrinsic metrics comprising data relating to the documents but not from the documents;
• and
• (viii) extracting the excerpts from the subset the documents, wherein the excerpts are extracted from the subset of documents based on the one or more criteria, the excerpts comprising text and non-text.