Code for ACL'2021 paper WARP 🌀 Word-level Adversarial ReProgramming

Last update: Nov 06, 2022

Overview

🌀 WARP: Word-level Adversarial ReProgramming

This repository contains code for ACL'2021 Paper WARP: Word-level Adversarial ReProgramming.

^{WARP adds a few trainable embeddings around the input, which causes the masked language model to predict the sentiment of the sentence in the SST-2 task.}

Transfer learning from pretrained language models recently became the dominant approach for solving many NLP tasks. A common approach to transfer learning for multiple tasks that maximize parameter sharing trains one or more task-specific layers on top of the language model.

In this paper, we present an alternative approach based on adversarial reprogramming, which extends earlier work on automatic prompt generation. Adversarial reprogramming attempts to learn task-specific word embeddings that, when concatenated to the input text, instruct the language model to solve the specified task.

Using up to 25K trainable parameters per task, this approach outperforms all existing methods that use up to 25M trainable parameters on the public leaderboard of the GLUE benchmark. Our method, initialized with task-specific human-readable prompts, also works in a few-shot setting, outperforming GPT-3 on two SuperGLUE tasks after training on just 32 samples.

Few-Shot Results

Set	Model	CB		RTE
Set	Model	F₁	Acc.	Acc.
dev
	GPT-3 Small	26.1	42.9	52.3
	GPT-3 Med	40.4	58.9	48.4
	GPT-3	57.2	82.1	72.9
	PET (ALBERT)	59.4	85.1	69.8
	iPET (ALBERT)	92.4	92.9	74.0
	WARP_init (ALBERT)	84.0	87.5	71.8
test
	GPT-3	52.0	75.6	69.0
	PET (ALBERT)	60.2	87.2	67.2
	iPET (ALBERT)	79.9	88.8	70.8
	WARP_init (ALBERT)	70.2	82.4	69.1

^{Results on SuperGLUE benchmark. The results for the test set are obtained from SuperGLUE evaluation server. We only show systems performing in a similar few-shot training setup using 32 examples.}

Setup

The code requires YerevaNN's internal version of allennlp

git clone https://github.com/YerevaNN/allennlp
git checkout warp
pip install .

Training

Linear Probing

for DATASET in 'cola' 'sst2' 'mrpc' 'qqp' 'stsb' 'mnli' 'rte' 'wnli' 'qnli'
do
    export HPARAMS='{
        "dataset": "'$DATASET'",
        "lr": 0.0001,
        "num_epochs": 20,
        "prompts": [],
        "reorder_optimized": false,
        "max_batch_size": 8,
        "max_tokens_sq": 262144, "on_logits":  false, "pooling_index":  null, "seed":  1}'
    python -m allennlp train \
    -s .aim/baseline-linear-${DATASET} configs/warp.jsonnet
done

WARP_0

"], "reorder_optimized": true, "max_batch_size": 8, "max_tokens_sq": 262144, "on_logits": "pre_decoder_layer_norm", "pooling_index": 1, "seed": 1 }' python -m allennlp train \ -s .aim/baseline-warp_0-${DATASET} configs/warp.jsonnet done ">

for DATASET in 'cola' 'sst2' 'mrpc' 'qqp' 'stsb' 'mnli' 'rte' 'wnli' 'qnli'
do
    export HPARAMS='{
        "dataset": "'$DATASET'",
        "lr": 0.0001,
        "num_epochs": 20,
        "prompts": [null, "
   
    "],
   
        "reorder_optimized": true,
        "max_batch_size": 8,
        "max_tokens_sq": 262144,
        "on_logits": "pre_decoder_layer_norm",
        "pooling_index": 1,
        "seed": 1
    }'
    python -m allennlp train \
    -s .aim/baseline-warp_0-${DATASET} configs/warp.jsonnet
done

Training WARP

", "prompts":[-10,-11,-12,-13,-14,null,-15,-16,-17,-18,-19,"
",-20,-21,-22,-23,-24,null,-25,-26,-27,-28,-29], "seed":1, "transformer_model":"roberta-large" }' python -m allennlp train \ -s .aim/t-${DATASET} configs/warp.jsonnet ">
export DATASET="rte"
export HPARAMS='{
    "benchmark":"super_glue",
    "classifier_init":null,
    "dataset":"'$DATASET'",
    "ensure_whitespace_between":false,
    "lr":0.001,
    "max_batch_size":8,
    "max_tokens_sq":262144,
    "num_epochs":30,
    "prompt_better_init":"
    
     ",
    
    "prompts":[-10,-11,-12,-13,-14,null,-15,-16,-17,-18,-19,"
    
     ",-20,-21,-22,-23,-24,null,-25,-26,-27,-28,-29],
    
    "seed":1,
    "transformer_model":"roberta-large"
}'
python -m allennlp train \
-s .aim/t-${DATASET} configs/warp.jsonnet

WARP_init

Few-Shot Experiments

", [-20, ","], null, [-29, "!"],-30,-31], "seed":3, "str_cut_frac":0, "transformer_model":"albert-xxlarge-v2", "validation_metric": null }' python -m allennlp train \ -s .aim/t-${DATASET}-`date +%s` configs/warp.jsonnet ">

export HPARAMS='{
    "benchmark":"super_glue",
    "classifier_init": {
        "entailment": " yes",
        "not_entailment": " instead"
    },
    "dataset":"few_rte",
    "eval_mode":false,
    "lr":0.001,
    "max_batch_size":2,
    "max_tokens_sq":131072,
    "num_epochs":100,
    "num_gradient_accumulation_steps":2,
    "prompt_better_init": "[PAD]",
    "prompts":[-10,-11,[-14,"\""],null,[-15,"\""],  [-16, "?"], "
   
    ", [-20, ","], null, [-29, "!"],-30,-31],
   
    "seed":3,
    "str_cut_frac":0,
    "transformer_model":"albert-xxlarge-v2",
    "validation_metric": null
}'
python -m allennlp train \
-s .aim/t-${DATASET}-`date +%s` configs/warp.jsonnet

",[-20,","],null,[-29,"!"],-30,-31], "seed":1, "str_cut_frac":0.06, "transformer_model":"albert-xxlarge-v2", "validation_metric":"+training_val_metric" }' python -m allennlp train \ -s .aim/t-${DATASET}-`date +%s` configs/warp.jsonnet ">

export HPARAMS='{
   "benchmark":"super_glue",
   "classifier_init":{
      "entailment":" yes",
      "not_entailment":" instead"
   },
   "dataset":"few_rte",
   "grad_norm":1,
   "lr":0.001,
   "max_batch_size":2,
   "max_tokens_sq":131072,
   "num_epochs":30,
   "num_gradient_accumulation_steps":2,
   "prompt_better_init":"[PAD]",
   "prompts":[-10,-11,[-14,"\""],null,[-15,"\""],[-16,"?"],"
   
    ",[-20,","],null,[-29,"!"],-30,-31],
   
   "seed":1,
   "str_cut_frac":0.06,
   "transformer_model":"albert-xxlarge-v2",
   "validation_metric":"+training_val_metric"
}'
python -m allennlp train \
-s .aim/t-${DATASET}-`date +%s` configs/warp.jsonnet

Evaluation

python -m allennlp predict \
  --silent --use-dataset-reader --cuda-device 0 \
  --batch-size 50 \
  --predictor glue --output-file v0.1/AX.tsv /data/arp/.aim/H-93ae5ae9 ax/test

python -m allennlp predict \
  --silent --use-dataset-reader --cuda-device 0 \
  --batch-size 50 \
  --predictor glue --output-file v0.1/MNLI-m.tsv /data/arp/.aim/H-93ae5ae9 test_matched

Citation

If you want to refer to our work use this bibTeX:

@inproceedings{hambardzumyan-etal-2021-warp,
    title = "{WARP}: {W}ord-level {A}dversarial {R}e{P}rogramming",
    author = "Hambardzumyan, Karen  and
      Khachatrian, Hrant  and
      May, Jonathan",
    booktitle = "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)",
    month = aug,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.acl-long.381",
    doi = "10.18653/v1/2021.acl-long.381",
    pages = "4921--4933"
}

Code for ACL'2021 paper WARP 🌀 Word-level Adversarial ReProgramming

Related tags

Overview

🌀 WARP: Word-level Adversarial ReProgramming

Few-Shot Results

Setup

Training

Linear Probing

WARP_0

Training WARP

WARP_init

Few-Shot Experiments

Evaluation

Citation

Owner

YerevaNN

Img-process-manual - Utilize Python Numpy and Matplotlib to realize OpenCV baisc image processing function

Universal Probability Distributions with Optimal Transport and Convex Optimization

Voice control for Garry's Mod

Implementation of CVPR 2021 paper "Spatially-invariant Style-codes Controlled Makeup Transfer"

Code for classifying international patents based on the text of their titles/abstracts

Official Implementation of CoSMo: Content-Style Modulation for Image Retrieval with Text Feedback

A clean and scalable template to kickstart your deep learning project 🚀 ⚡ 🔥

YoHa - A practical hand tracking engine.

DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism (SVS & TTS); AAAI 2022; Official code

TorchGRL is the source code for our paper Graph Convolution-Based Deep Reinforcement Learning for Multi-Agent Decision-Making in Mixed Traffic Environments for IV 2022.

Monk is a low code Deep Learning tool and a unified wrapper for Computer Vision.

Code for reproducing our paper: LMSOC: An Approach for Socially Sensitive Pretraining

Monocular Depth Estimation Using Laplacian Pyramid-Based Depth Residuals

PyTorch Implementation of Daft-Exprt: Robust Prosody Transfer Across Speakers for Expressive Speech Synthesis

Code for: https://berkeleyautomation.github.io/bags/

商品推荐系统

The dataset and source code for our paper: "Did You Ask a Good Question? A Cross-Domain Question IntentionClassification Benchmark for Text-to-SQL"

POPPY (Physical Optics Propagation in Python) is a Python package that simulates physical optical propagation including diffraction

General-purpose program synthesiser

FinEAS: Financial Embedding Analysis of Sentiment 📈