The terms that appear in many more new papers this week than in the eight weeks before, from several independent teams, grouped into topics and explained in plain language.
×2.3 compared with usual · 40 papers from 39 teams
Up week after week, not a one-week event.
These studies explore AI systems designed for specific decisions, including assessing clinical-trial criteria and ranking information retrieved for an answer. The increase may reflect growing interest in testing specialized models and keeping authoritative rules separate from potentially untrustworthy evidence.
Correcting systematic errors in statistical estimates
×4.3 compared with usual · 12 papers from 12 teams
Already this frequent the week of 3 November 2025, a year ago: perhaps a yearly cycle (a conference deadline, a season, an observing campaign).
Statistical estimates can systematically miss their targets, even when the calculations are performed correctly. These papers study how to measure and correct such bias in settings including panel data and information estimation; the rise may reflect parallel efforts to make uncertainty estimates and conclusions more reliable.
Tracking charged particles and understanding their interactions
×3.6 compared with usual · 14 papers from 13 teams
Already this frequent the week of 6 April 2026.
Charged particles provide measurable traces of collisions and of energy deposited in matter. The examples cover particle detectors, collision measurements and simulations for radiation treatment, so the increase appears to span several applications rather than one shared development.
Each week, we count how many papers submitted that week (first versions) use each term of one to three words in their title and abstract; acronyms such as RAG or JWST count, and plurals with their singular. A term rises when its share of papers is at least twice its share over the eight weeks before, when so large a rise is very unlikely by chance (a log-likelihood ratio of at least 15), and when it comes from at least three teams, none signing more than half of the papers: one group posting a series is not a trend. An AI then groups the rising terms into topics and names them; it never decides what rises: the counts do.
New: the term was almost absent from these papers before.
Rising for several weeks: up three weeks in a row or more, not a one-week event.
Recurring: as frequent at this time last year: perhaps a yearly cycle, such as a conference deadline.
Comeback: as frequent some months ago, then less used.
Sudden rise: new this week: often an event, a release, a data set or a deadline.
The curve on each topic is the share of papers using its main term, per 10,000 papers, over the last six months. Counting the previous Monday-to-Sunday submission week starts on Tuesday at 06:00 UTC, after the normal Monday arXiv announcement. Delayed announcements may still be missing from the first count.
