News
30 Oct 2024
Introducing Adarga’s next-gen search: enhanced information retrieval across in-house and open-source data
Introduction: Increasing speed, precision, and context
Vantage, Adarga’s specialist AI software delivering decision advantage to organisations worldwide, has been further enhanced with the introduction of next-gen search.
Representing the latest in Adarga’s leading information intelligence capabilities, next-gen search provides analysts, researchers, and planners with highly relevant, domain-specific results to their most complex queries, at increased speed and with increased precision. This is a vital capability in the overwhelming data environment in which they operate.
Most importantly, next-gen search provides a level of context that drives new levels of efficiencies in the analyst workflow, building on our long-standing ability to provide a single, secure platform in which users can interrogate both their internal knowledge bases and open-source information to surface the most robust results.
This contextualisation is rooted in the specialist security needs of our customers in the public and private sector – such as 14 North who are using Vantage to augment geopolitical risk intelligence on Sub-Saharan Africa, and UK Strategic Command’s Defence Support who are using Vantage to analyse threats to UK military operations worldwide.
What is Adarga’s next-gen search?
Adarga’s next-gen search is an advanced information retrieval system embedded within Adarga Vantage - delivering answers that go beyond simple keyword matching to include related concepts, ideas, and connections.
It leverages a hybrid search model, combining the precision of traditional keyword matching with the depth of vector embeddings. This dual approach enables the system not only to search for specific terms but also to understand their broader meanings through semantic search, an advanced meaning-based method using natural language processing.
These latest capabilities have shown a 25% improvement in search accuracy compared to our previous methods.
What does next-gen search cover: in-house, open-source data, and more.
Adarga Vantage has been designed for customers who operate in high-stakes environments and need to contend with fast-paced global dynamics from operational, strategic, or tactical perspectives. Its data environment is built to reflect this, providing a single, secure platform that can be curated for each customer. It provides constantly updated:
- Highly curated publicly available information - sourced by our team of geopolitical experts from 100+ countries, and automatically translated from 75+ languages. This includes 10m+ assets sitting in our baseline package, growing at 20k+ assets per day.
- Government and academic reports.
- Near real-time events sourced from social media and mapped to a location.
- Internal proprietary data uploaded to Adarga Vantage such as humint reports.
- Commercially available data.
Why is in-house and open-source data layering key for robust analysis outputs?
The ability to return rapid, relevant results on a query is vital. But the ability to return results like this, rooted in the context of your own business and internal knowledge-base, takes analysis speed and quality to the next level. It provides a unified picture of information within which you can validate and corroborate diverse insight, challenge findings, and uncover internal knowledge on a topic you didn’t even know existed.
- As a private sector company operating in risk and security, this means the ability to search, explore, and automatically connect insights from your internal human intelligence reports or risk reports with emerging open-source information to get a full view of what’s going on in the ground in a region of interest.
- As a defence/national security company, this means the ability to incorporate historic lessons learnt, buried in documentation such as post operational tour reports or intelligence reports from many years gone, into current planning.
We use Vantage to securely fuse insights from past reports with emerging open-source data. It’s a powerful combination that saves us time, driving efficiency, productivity, and quality of outputs.
- William Linder, 14 North’s CEO
Going beyond search: related query suggestion, question and answers, and saved searches
One of the standout features of next-gen search is its ability to suggest related queries and allow users to refine their searches as they navigate the results. For example, after reaching the top 50 results, users can adjust their search criteria, explore related searches, or use the Q&A feature for further guided question suggestions.
Q&A enables users to ask questions of their own curated data sets and reports. It generates well-evidenced, natural language answers using our own carefully calibrated, rigorously tested generative AI model - with robust guardrails in place.
The asset panel on the right of search (shown in the first video) enables an easy deep-dive into the results, with AI-generated summaries providing an understanding of the contents 90% faster; translated content and original source documents; video transcripts; identified events; and relationships between people, places and organisations that have been identified within the material.
The platform also provides a range of filtering options to fine-tune search results, including:
- Time and date filters
- Country of origin filters to identify different perspectives
- Source-specific filters to assess and compare government-only sources, think tanks, news outlets, or internal knowledge bases
- Advanced criteria filters, allowing Boolean-only searches for precision-based queries. (Boolean is a logic-based approach to search using standard operators such as “AND,” “OR” or “NOT.”)
The ‘saved searches’ capability allows users to set up a bank of relevant searches at a project level, enabling them to monitor topics of interest with fast-moving dynamics without starting-a-new each time.
Providing domain-specific results
Finally, given our models are trained on domain-specific data from the military, security, and geopolitical sphere, Vantage has been geared up from all angles to deliver the most relevant results to users. As domain experts, having an understanding of our customers’ challenges and operational environments is fundamental to ensuring Adarga Vantage can meet their most critical analysis needs.
Delivering real-world Impact
A senior in-house geopolitical analyst at Adarga described next-gen search as a "gamechanger” in her workflow. "I was recently working on a project related to China’s influence in cobalt mining in Africa, and next-gen search not only delivered rapid, highly relevant results but also guided me to areas of the topic I wouldn’t have otherwise found. Through suggested searches, I was able to quickly build a detailed picture - not only of China's influence but also of the human rights issues in the Democratic Republic of Congo."
Staying at the forefront of innovation
Ensuring we remain at the forefront of innovation when it comes to information retrieval, and augmenting human intelligence in all facets of information analysis, lies at the core of what we do. It’s vital for the needs of our specialist customer-base who operate in high-stakes environments and require a solution designed to help them stay ahead of fast-moving global dynamics, striking the right balance of human machine teaming.
For analysts, time is of the essence. Getting them to the most relevant information, quicker allows them to focus on higher-value activities and apply more time to critical thinking using their specialist skills - enabling the delivery of a better breadth, depth and quality of insight/intelligence to decision-makers for strategic advantage.
We've outlined some examples of customer applications of Vantage below:
- Augmenting intelligence analysis
- Analysing threats
- Enriching security and risk assessments
- Enhancing scenario planning
- Accelerating policy research
- Monitoring safety of people on the ground
To find out more, get in touch.
