Food Product Tracing Technology Capabilities and Interoperability
Bibliographic record
Abstract
Despite the best efforts of food safety and food defense professionals, contaminated food continues to enter the food supply. It is imperative that contaminated food be removed from the supply chain as quickly as possible to protect public health and stabilize markets. To solve this problem, scores of technology companies purport to have the most effective, economical product tracing system. This study sought to compare and contrast the effectiveness of these systems at analyzing product tracing information to identify the contaminated ingredient and likely source, as well as distribution of the product. It also determined if these systems can work together to better secure the food supply (their interoperability). Institute of Food Technologists (IFT) hypothesized that when technology providers are given a full set of supply-chain data, even for a multi-ingredient product, their systems will generally be able to trace a contaminated product forward and backward through the supply chain. However, when provided with only a portion of supply-chain data, even for a product with a straightforward supply chain, it was expected that interoperability of the systems will be lacking and that there will be difficulty collaborating to identify sources and/or recipients of potentially contaminated product. IFT provided supply-chain data for one complex product to 9 product tracing technology providers, and then compared and contrasted their effectiveness at analyzing product tracing information to identify the contaminated ingredient and likely source, as well as distribution of the product. A vertically integrated foodservice restaurant agreed to work with IFT to secure data from its supply chain for both a multi-ingredient and a simpler product. Potential multi-ingredient products considered included canned tuna, supreme pizza, and beef tacos. IFT ensured that all supply-chain data collected did not include any proprietary information or information that would otherwise identify the supply-chain partner who provided the information prior to sharing this information with product tracing technology providers. The 9 traceability solution providers who agreed to participate in this project have their systems deployed in a wide range of sectors within the food industry including, but not limited to, livestock, dairy, produce, fruits, seafood, meat, and pork; as well as in pharmaceutical, automotive, retail, and other industries. Some have also been implemented across the globe including Canada, China, USA, Norway, and the EU, among others. This broad commercial use ensures that the findings of this work are applicable to a broad spectrum of the food system. Six of the 9 participants successfully completed the data entry phase of this test. To verify successful data entry for these 6, a demo or screenshots of the data set from each system's user interface was requested. Only 4 of the 6 were able to provide us with this evidence for verification. Of the 6 that completed data entry and moved on to the scenarios phase of the test, 5 were able to provide us with the responses to the scenarios. Time metrics were useful for evaluating the scalability and usability of each technology. Scalability was derived from the time it took to enter the nonstandardized data set into the system (ranges from 7 to 11 d). Usability was derived from the time it took to query the scenarios and provide the results (from a few hours to a week). The time was measured in days it took for the participants to respond after we supplied them all the information they would need to successfully execute each test/scenario. Two of the technology solution providers successfully implemented and participated in a proof-of-concept interoperable framework during Year 2 of this study. While not required, they also demonstrated this interoperability capability on the FSMA-mandated food product tracing pilots for the U.S. FDA. This has significant real-world impact since the demonstration of interoperability enables U.S. FDA to obtain evidence on the importance and impact of data-sharing moving forward. Another real-world accomplishment is the modification or upgrade of commercial technology solutions to enhance or implement interoperability. As these systems get deployed by clients in the food industry, interoperability will no longer be an afterthought but will be built into their traceability systems. In turn, industry and regulators will better understand the capabilities of the currently available technologies, and the technology provider community will identify ways in which their systems may be further developed to increase interoperability and utility.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.035 | 0.061 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.007 | 0.004 |
| Science and technology studies | 0.002 | 0.004 |
| Scholarly communication | 0.013 | 0.022 |
| Open science | 0.003 | 0.010 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".