Papers › An Effective Domain Adaptive Post-Training Method for BERT in Response Selection

An Effective Domain Adaptive Post-Training Method for BERT in Response Selection

13 Aug 2019arXiv:1908.04812archive 2025-07-28

Taesun Whang, Dongyub Lee, Chanhee Lee, Kisu Yang, Dongsuk Oh, Heuiseok Lim

We focus on multi-turn response selection in a retrieval-based dialog system. In this paper, we utilize the powerful pre-trained language model Bi-directional Encoder Representations from Transformer (BERT) for a multi-turn dialog system and propose a highly effective post-training method on domain-specific corpus. Although BERT is easily adopted to various NLP tasks and outperforms previous baselines of each task, it still has limitations if a task corpus is too focused on a certain domain. Post-training on domain-specific corpus (e.g., Ubuntu Corpus) helps the model to train contextualized representations and words that do not appear in general corpus (e.g., English Wikipedia). Experimental results show that our approach achieves new state-of-the-art on two response selection benchmarks (i.e., Ubuntu Corpus V1, Advising Corpus) performance improvement by 5.9% and 6% on R@1.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

taesunwhang/BERT-ResSel mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Conversational Response SelectionLanguage ModelingLanguage ModellingRetrieval

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Conversational Response Selection Douban BERT MAP 0.591 #10 of 16 Archive leaderboard report
Conversational Response Selection Douban BERT MRR 0.633 #10 of 16 Archive leaderboard report
Conversational Response Selection Douban BERT P@1 0.454 #10 of 16 Archive leaderboard report
Conversational Response Selection Douban BERT R10@1 0.280 #10 of 16 Archive leaderboard report
Conversational Response Selection Douban BERT R10@2 0.470 #10 of 16 Archive leaderboard report
Conversational Response Selection Douban BERT R10@5 0.828 #10 of 16 Archive leaderboard report
Conversational Response Selection RRS BERT MAP 0.625 #4 of 7 Archive leaderboard report
Conversational Response Selection RRS BERT MRR 0.639 #4 of 7 Archive leaderboard report
Conversational Response Selection RRS BERT P@1 0.453 #4 of 7 Archive leaderboard report
Conversational Response Selection RRS BERT R10@1 0.404 #4 of 7 Archive leaderboard report
Conversational Response Selection RRS BERT R10@2 0.606 #4 of 7 Archive leaderboard report
Conversational Response Selection RRS BERT R10@5 0.875 #4 of 7 Archive leaderboard report
Conversational Response Selection RRS Ranking Test BERT NDCG@3 0.625 #3 of 4 Archive leaderboard report
Conversational Response Selection RRS Ranking Test BERT NDCG@5 0.714 #3 of 4 Archive leaderboard report
Conversational Response Selection Ubuntu Dialogue (v1, Ranking) BERT-VFT R10@1 0.855 #10 of 25 Archive leaderboard report
Conversational Response Selection Ubuntu Dialogue (v1, Ranking) BERT-VFT R10@2 0.928 #10 of 25 Archive leaderboard report
Conversational Response Selection Ubuntu Dialogue (v1, Ranking) BERT-VFT R10@5 0.985 #10 of 25 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Absolute Position EncodingsAdamAttentionAttention DropoutBERTBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerLinear Warmup With Linear DecayMulti-Head AttentionPosition-Wise Feed-Forward LayerReLUResidual ConnectionSoftmaxTransformerWeight DecayWordPiece

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections