perturb_for_table / table_result /2407.00091v1_output.json
wcy
'modify'
0803c45
raw
history blame
7.67 kB
[
{
"path": "table_paper/2407.00091v1.json",
"table_id": "1",
"section": "3.1",
"all_context": [
"In the first phase, we study the effects of the parameter through offline simulation of the search system.",
"Table 1 summarizes the aggregate statistics of map-results corresponding to different values of .",
"The lower bound of the exploration is set by product experience considerations, and the upper bound the result of diminishing effects.",
"The baseline for all these comparisons is map-results with no filtering and a fixed limit of pins.",
"Table 1 validates that the Bookability Filter is having the intended effect on map-results.",
"The second phase of our testing investigates how users react to various values of .",
"We run multiple A/B experiments online, where control applies no filtering on map-results, and treatments apply the Bookability Filter with different values of .",
"Table 2 summarizes the effect of on searchers.",
"A brief explanation of the key metrics evaluated in the online experiments: Uncanceled bookings: This is the top line metric, the number of bookings made by searchers that were not cancelled.",
"5-star trips: Trips booked by searchers that resulted in 5-star rating after checkout, evaluated days after end of experiment.",
"Average impressions to discovery: The average number of distinct search results that a booker saw before clicking the listing that was booked.",
"This measures the cognitive load of making a booking.",
"Average clicks to discovery: The number of distinct search results clicked by a booker before clicking the listing that was booked.",
"This is an alternative measure of the cognitive load of making a booking.",
"In the final phase, we fix the value of to corresponding to maximum user benefit, and repeat the online A/B experiment at a larger scale, allocating million searchers worldwide to each of control and treatment.",
"This grinds the p-value of the key metrics below .",
"Uncanceled bookings increase by , measured as a percentage of overall global bookings at Airbnb, making it one of the largest improvements launched over the last several years.",
"5-star trips increase by indicating not only growth in bookings, but a growth in quality bookings.",
"Average number of results seen by the searcher before clicking on the booked listing reduces by , while search results clicked drop by .",
"The reduction in effort to locate the booked listing, due to removal of inferior choices, is the key mechanism driving the gain in bookings.",
""
],
"target_context_ids": [
1,
4,
5
],
"selected_paragraphs": [
"[paragraph id = 1] Table 1 summarizes the aggregate statistics of map-results corresponding to different values of .",
"[paragraph id = 4] Table 1 validates that the Bookability Filter is having the intended effect on map-results.",
"[paragraph id = 5] The second phase of our testing investigates how users react to various values of ."
],
"table_html": "<figure class=\"ltx_table\" id=\"S3.T1\">\n<table class=\"ltx_tabular ltx_centering ltx_guessed_headers ltx_align_middle\" id=\"S3.T1.25\">\n<thead class=\"ltx_thead\">\n<tr class=\"ltx_tr\" id=\"S3.T1.5.5\">\n<th class=\"ltx_td ltx_align_left ltx_th ltx_th_column ltx_border_tt\" id=\"S3.T1.1.1.1\"></th>\n<th class=\"ltx_td ltx_align_left ltx_th ltx_th_column ltx_border_tt\" id=\"S3.T1.2.2.2\"></th>\n<th class=\"ltx_td ltx_align_left ltx_th ltx_th_column ltx_border_tt\" id=\"S3.T1.3.3.3\"></th>\n<th class=\"ltx_td ltx_align_left ltx_th ltx_th_column ltx_border_tt\" id=\"S3.T1.4.4.4\"></th>\n<th class=\"ltx_td ltx_align_left ltx_th ltx_th_column ltx_border_tt\" id=\"S3.T1.5.5.5\"></th>\n</tr>\n</thead>\n<tbody class=\"ltx_tbody\">\n<tr class=\"ltx_tr\" id=\"S3.T1.9.9\">\n<td class=\"ltx_td ltx_align_left ltx_border_t\" id=\"S3.T1.9.9.5\">Number of map pins</td>\n<td class=\"ltx_td ltx_align_left ltx_border_t\" id=\"S3.T1.6.6.1\"></td>\n<td class=\"ltx_td ltx_align_left ltx_border_t\" id=\"S3.T1.7.7.2\"></td>\n<td class=\"ltx_td ltx_align_left ltx_border_t\" id=\"S3.T1.8.8.3\"></td>\n<td class=\"ltx_td ltx_align_left ltx_border_t\" id=\"S3.T1.9.9.4\"></td>\n</tr>\n<tr class=\"ltx_tr\" id=\"S3.T1.13.13\">\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.13.13.5\">Average booking probability</td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.10.10.1\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.11.11.2\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.12.12.3\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.13.13.4\"></td>\n</tr>\n<tr class=\"ltx_tr\" id=\"S3.T1.17.17\">\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.17.17.5\">Average total price</td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.14.14.1\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.15.15.2\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.16.16.3\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.17.17.4\"></td>\n</tr>\n<tr class=\"ltx_tr\" id=\"S3.T1.21.21\">\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.21.21.5\">Average number of reviews</td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.18.18.1\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.19.19.2\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.20.20.3\"></td>\n<td class=\"ltx_td ltx_align_left\" id=\"S3.T1.21.21.4\"></td>\n</tr>\n<tr class=\"ltx_tr\" id=\"S3.T1.25.25\">\n<td class=\"ltx_td ltx_align_left ltx_border_bb\" id=\"S3.T1.25.25.5\">Average review rating</td>\n<td class=\"ltx_td ltx_align_left ltx_border_bb\" id=\"S3.T1.22.22.1\"></td>\n<td class=\"ltx_td ltx_align_left ltx_border_bb\" id=\"S3.T1.23.23.2\"></td>\n<td class=\"ltx_td ltx_align_left ltx_border_bb\" id=\"S3.T1.24.24.3\"></td>\n<td class=\"ltx_td ltx_align_left ltx_border_bb\" id=\"S3.T1.25.25.4\"></td>\n</tr>\n</tbody>\n</table>\n<figcaption class=\"ltx_caption ltx_centering\"><span class=\"ltx_tag ltx_tag_table\"><span class=\"ltx_text\" id=\"S3.T1.31.3.1\" style=\"font-size:90%;\">Table 1</span>. </span><span class=\"ltx_text\" id=\"S3.T1.29.2\" style=\"font-size:90%;\">Offline exploration of compared against a baseline with no filtering, which is conceptually equivalent to .</span></figcaption>\n</figure>",
"perturb_sentence_id": [
1,
2
],
"output": {
"perturbed_statement": "[paragraph id = 1] Table 1 summarizes the effect of the Bookability Filter on different user demographics. The lower bound of the exploration is set by product experience considerations, and the upper bound the result of diminishing effects.",
"perturbed_explanation": "1. The original explanation states that Table 1 summarizes the aggregate statistics of map-results corresponding to different values of an unspecified variable. The lower bound of exploration is determined by product experience considerations, while the upper bound is due to diminishing effects. 2. The statement is incorrect because it claims that Table 1 summarizes the effect of the Bookability Filter on different user demographics, which is not mentioned in the original context. Instead, the context indicates that Table 1 validates the effect of the Bookability Filter on map-results, not on user demographics."
}
}
]