Query expansion has been proved to be effective in improving recall and precision of first-stage retrievers, and yet its influence on a complicated, state-of-the-art cross-encoder ranker remains under-explored. We first show that directly applying the expansion techniques in the current literature to state-of-the-art neural rankers can result in deteriorated zero-shot performance. To this end, we propose GFF, a pipeline that includes a large language model and a neural ranker, to Generate, Filter, and Fuse query expansions more effectively in order to improve the zero-shot ranking metrics such as nDCG@10. Specifically, GFF first calls an instruction-following language model to generate query-related keywords through a reasoning chain. Leveraging self-consistency and reciprocal rank weighting, GFF further filters and combines the ranking results of each expanded query dynamically. By utilizing this pipeline, we show that GFF can improve the zero-shot nDCG@10 on BEIR and TREC DL 2019/2020. We also analyze different modelling choices in the GFF pipeline and shed light on the future directions in query expansion for zero-shot neural rankers.

They used to be little more than buzzwords, but that's not the case anymore. Blockchain and artificial intelligence (AI) have evolved into leading technologies that power innovation across almost every industry. From blockchain's ability to track-and-trace food supply chains to the integration of AI in almost every facet of healthcare, each is tackling some of the world's biggest challenges. It turns out they actually work well in unison, too, building and organizing immense databases, strengthening cybersecurity protocols and performing tasks in a fraction of the time it takes humans. Here are 30 companies that are combining blockchain and AI with great results.

CASTRO VALLEY, CALIFORNIA – A man was killed when the Tesla automobile he was driving veered off a road, crashed through a fence and plunged into a pond, authorities said Monday. California Highway Patrol spokesman Daniel Jacowitz said rescuers pulled the Tesla Model S from the pond early Monday and found the man's body inside. The driver was identified as Keith Leung, 34, of Danville, California, said Sgt. Ray Kelly, spokesman for the Alameda County Sheriff's office. Kelly said it was too soon to know if the vehicle's semi-autonomous Autopilot mode was engaged when the crash occurred or whether the driver may have been speeding or intoxicated. Photographs of the car show that its back-end was destroyed, its hood crumpled and windows shattered.

Riding in a self-driving car involves a certain level of risk taking, especially in these very early days of the technology. Recognizing the need to build trust among normal people, Waymo, the self-driving unit of Google-parent Alphabet, announced it would be partnering with an insurance startup to cover riders in its soon-to-be-released driverless ride-hailing service. Trov, a five-year-old insurance tech startup based in Danville, Calif., said it would work with Waymo to insure passengers for lost and damaged property and trip-related medical expenses. In other words, if your driverless Waymo is involved in a fender bender -- or, god forbid, something worse -- your robot-induced whiplash treatment will be covered. Importantly, passengers won't have to pay for the coverage, nor will they know that Trov is the insurer.