We lifted 7 functions out of this paper's own repositories and ran 2 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| HsuWanTing/unified-summarization | pwc_unofficial | 2 of 7 |
| Function | Status | Where it lives |
|---|---|---|
| make_html_safe | Ran | HsuWanTing/unified-summarization/end2end/evaluate.py code served (permissive licence) · get_code("84fcb74e05d73aff") |
| sort_hyps | Ran | HsuWanTing/unified-summarization/end2end/beam_search.py code served (permissive licence) · get_code("8891bb52fbe37510") |
| attention_decoder_one_step | Not yet run | HsuWanTing/unified-summarization/rewriter/attention_decoder.py code served (permissive licence) · get_code("77220f0cda583943") |
| calc_running_avg_loss | Not yet run | HsuWanTing/unified-summarization/util.py code served (permissive licence) · get_code("3fb8a5469710fb6f") |
| get_select_accuracy_one_thres | Not yet run | HsuWanTing/unified-summarization/util.py code served (permissive licence) · get_code("0325f9d740a777ae") |
| linear | Not yet run | HsuWanTing/unified-summarization/rewriter/attention_decoder.py code served (permissive licence) · get_code("75bf8deb2b42c686") |
| rouge_log | Not yet run | HsuWanTing/unified-summarization/end2end/evaluate.py code served (permissive licence) · get_code("7754f6d79787fe92") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
We propose a unified model combining the strength of extractive and abstractive summarization. On the one hand, a simple extractive model can obtain sentence-level attention with high ROUGE scores but less readable. On the other hand, a more complicated abstractive model can obtain word-level dynamic attention to generate a more readable paragraph. In our model, sentence-level attention is used to modulate the word-level attention such that words in less attended sentences are less likely to be generated. Moreover, a novel inconsistency loss function is introduced to penalize the inconsistency between two levels of attentions. By end-to-end training our model with the inconsistency loss and original losses of extractive and abstractive models, we achieve state-of-the-art ROUGE scores while being the most informative and readable summarization on the CNN/Daily Mail dataset in a solid human evaluation.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("1805.06266")
get_code_for_paper("1805.06266")
have("1805.06266")
Connect an agent — have() is free.