Automated Feed Identifier (AFI)

Stage in GSE Lifecyle

Primary stage : Maintain

Related stage(s) : Builds Capability | Define/Design | Evaluate

The Official Control Regulation (OCR) 2019/625 requires competent authorities to perform official controls regularly, on a risk basis and with appropriate frequency, on feed entering the EU. There currently is variation between local authorities in how feed consignments are identified but most authorities will manually examine a manifest for a feed declaration.

Strategic Surveillance’s AFI uses the Machine Learning technology of Optical Character Recognition to read multiple formats of electronic documents. The tool’s text extraction engine transforms pixels from the images or pdfs into readable text. The text found is then compared, using another Machine Learning technology called NLP (Natural Language Processing), to various feed catalogues to identify feed commodities listed in the manifests.

The tool is expected to drastically reduce the amount of manual effort involved and is currently being tested by two Local Authorities. It will also be extended to identifying food commodities in manifests.

An image of a stack of files.

---

Team

Strategic Surveillance

---

Department(s)

Food Standards Agency

Department for the Environment, Food and Rural Affairs (DEFRA)

---

Contact Details

Email : strategic.surveillance@food.gov.uk

---

Description

The Strategic Surveillance Service is a data science team formed in 2017 to strengthen the FSA’s food surveillance arsenal and support its food safety mission. The team develops tools and techniques to turn data into intelligence, using machine learning and artificial intelligence. This helps us and our external users make quicker, better informed actions to protect consumers.