bert-base-uncased-job-bias-seq-cls

Model Card for bert-base-uncased-job-bias-seq-cls

<!-- Provide a quick summary of what the model is/does. -->

Model Details

<!-- Provide a longer summary of what this model is. -->

The model is a multi-label classifier designed to detect various types of bias within job descriptions.

• Developed by: Tristan Everitt and Paul Ryan
• Model type: Encoder
• Language(s) (NLP): en
• License: apache-2.0
• Finetuned from model: google-bert/bert-base-uncased

Model Sources

• Repository: https://github.com/2024-mcm-everitt-ryan
• Paper: In Progress

Uses

The primary target audience for these models are researchers dedicated to identifying biased language in job descriptions.

Out-of-Scope Use

Due to the limitations inherent in large-scale language models, they should not be utilised in applications requiring factual or accurate outputs. These models do not distinguish between fact and fiction, and implicit biases are inherently subjective.

Moreover, as language models mirror the biases present in their training data, they should not be deployed in systems that directly interact with humans unless the deployers have first conducted a thorough analysis of relevant biases for the specific use case.

Bias, Risks, and Limitations

It is imperative that all users, both direct and downstream, are aware of the risks, biases, and limitations associated with this model. Important considerations include:

• Bias in Training Data: The model may inherit and perpetuate biases from the data it was trained on.
• Subjectivity of Bias: Bias detection is inherently subjective, and perceptions of bias can differ across contexts and users.
• Accuracy Concerns: The model’s outputs are not guaranteed to be true or accurate, making it unsuitable for applications that require reliable information.
• Human Interaction Risks: When incorporated into systems that interact with humans, the model’s biases may affect interactions and decision-making, potentially leading to unintended consequences.

It is crucial for users to conduct comprehensive evaluations and consider these factors when applying the model in any context.

How to Get Started with the Model

Use the code below to get started with the model.

```python
from transformers import pipeline

pipe = pipeline("text-classification", model="2024-mcm-everitt-ryan/bert-base-uncased-job-bias-seq-cls", return_all_scores=True)

results = pipe("Join our dynamic and fast-paced team as a Junior Marketing Specialist. We seek a tech-savvy and energetic individual who thrives in a vibrant environment. Ideal candidates are digital natives with a fresh perspective, ready to adapt quickly to new trends. You should have recent experience in social media strategies and a strong understanding of current digital marketing tools. We're looking for someone with a youthful mindset, eager to bring innovative ideas to our young and ambitious team. If you're a recent graduate or early in your career, this opportunity is perfect for you!")
print(results)
```
>> [[
{'label': 'age', 'score': 0.9883460402488708}, 
{'label': 'disability', 'score': 0.00787709467113018}, 
{'label': 'feminine', 'score': 0.007224376779049635}, 
{'label': 'general', 'score': 0.09967829287052155}, 
{'label': 'masculine', 'score': 0.0035264550242573023}, 
{'label': 'racial', 'score': 0.014618005603551865}, 
{'label': 'sexuality', 'score': 0.005568435415625572}
]]

Training Details

Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

• 2024-mcm-everitt-ryan/benchmark

Training Hyperparameters

accelerator_config="{'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}", adafactor=false, adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, auto_find_batch_size=false, batch_eval_metrics=false, bf16=false, bf16_full_eval=false, data_seed="None", dataloader_drop_last=false, dataloader_num_workers=0, dataloader_persistent_workers=false, dataloader_pin_memory=true, dataloader_prefetch_factor="None", ddp_backend="None", ddp_broadcast_buffers="None", ddp_bucket_cap_mb="None", ddp_find_unused_parameters="None", ddp_timeout=1800, deepspeed="None", disable_tqdm=false, dispatch_batches="None", do_eval=true, do_predict=false, do_train=false, eval_accumulation_steps="None", eval_batch_size=8, eval_delay=0, eval_do_concat_batches=true, eval_on_start=false, eval_steps="None", eval_strategy="epoch", evaluation_strategy="None", fp16=false, fp16_backend="auto", fp16_full_eval=false, fp16_opt_level="O1", fsdp="[]", fsdp_config="{'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}", fsdp_min_num_params=0, fsdp_transformer_layer_cls_to_wrap="None", full_determinism=false, gradient_accumulation_steps=1, gradient_checkpointing="(False,)", gradient_checkpointing_kwargs="None", greater_is_better=false, group_by_length=true, half_precision_backend="auto", ignore_data_skip=false, include_inputs_for_metrics=false, jit_mode_eval=false, label_names="None", label_smoothing_factor=0.0, learning_rate=3e-05, length_column_name="length", load_best_model_at_end=true, local_rank=0, lr_scheduler_kwargs="{}", lr_scheduler_type="linear", max_grad_norm=1.0, max_steps=-1, metric_for_best_model="loss", mp_parameters="", neftune_noise_alpha="None", no_cuda=false, num_train_epochs=3, optim="adamw_torch", optim_args="None", optim_target_modules="None", past_index=-1, per_device_eval_batch_size=8, per_device_train_batch_size=8, per_gpu_eval_batch_size="None", per_gpu_train_batch_size="None", prediction_loss_only=false, ray_scope="last", remove_unused_columns=true, report_to="[]", restore_callback_states_from_checkpoint=false, resume_from_checkpoint="None", seed=42, skip_memory_metrics=true, split_batches="None", tf32="None", torch_compile=false, torch_compile_backend="None", torch_compile_mode="None", torchdynamo="None", tpu_num_cores="None", train_batch_size=8, use_cpu=false, use_ipex=false, use_legacy_prediction_loop=false, use_mps_device=false, warmup_ratio=0.0, warmup_steps=0, weight_decay=0.001 <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->

Results

              precision    recall  f1-score   support

         age       0.84      0.54      0.66        80
  disability       0.79      0.41      0.54        80
    feminine       0.95      0.94      0.94        80
     general       0.90      0.24      0.38        80
   masculine       0.72      0.54      0.61        80
     neutral       0.32      0.86      0.47        80
      racial       0.89      0.81      0.85        80
   sexuality       0.94      0.81      0.87        80

   micro avg       0.67      0.64      0.66       640
   macro avg       0.79      0.64      0.66       640
weighted avg       0.79      0.64      0.66       640
 samples avg       0.67      0.69      0.67       640

Environmental Impact

<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->

Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).

• Hardware Type: 1 X NVIDIA L40S
• Hours used: 0.12
• Cloud Provider: N/A
• Compute Region: N/A
• Carbon Emitted: N/A

Compute Infrastructure

• Linux 6.5.0-35-generic x86_64
• MemTotal: 1056613768 kB
• 256 X AMD EPYC 7702 64-Core Processor
• GPU_0: NVIDIA L40S

Software

python 3.10.12, accelerate 0.32.1, aiohttp 3.9.5, aiosignal 1.3.1, anyio 4.2.0, argon2-cffi 23.1.0, argon2-cffi-bindings 21.2.0, arrow 1.3.0, asttokens 2.4.1, async-lru 2.0.4, async-timeout 4.0.3, attrs 23.2.0, awscli 1.33.26, Babel 2.14.0, beautifulsoup4 4.12.3, bitsandbytes 0.43.1, bleach 6.1.0, blinker 1.4, botocore 1.34.144, certifi 2024.2.2, cffi 1.16.0, charset-normalizer 3.3.2, click 8.1.7, cloudpickle 3.0.0, colorama 0.4.6, comm 0.2.1, cryptography 3.4.8, dask 2024.7.0, datasets 2.20.0, dbus-python 1.2.18, debugpy 1.8.0, decorator 5.1.1, defusedxml 0.7.1, dill 0.3.8, distro 1.7.0, docutils 0.16, einops 0.8.0, entrypoints 0.4, evaluate 0.4.2, exceptiongroup 1.2.0, executing 2.0.1, fastjsonschema 2.19.1, filelock 3.13.1, flash-attn 2.6.1, fqdn 1.5.1, frozenlist 1.4.1, fsspec 2024.2.0, h11 0.14.0, hf_transfer 0.1.6, httpcore 1.0.2, httplib2 0.20.2, httpx 0.26.0, huggingface-hub 0.23.4, idna 3.6, importlib_metadata 8.0.0, iniconfig 2.0.0, ipykernel 6.29.0, ipython 8.21.0, ipython-genutils 0.2.0, ipywidgets 8.1.1, isoduration 20.11.0, jedi 0.19.1, jeepney 0.7.1, Jinja2 3.1.3, jmespath 1.0.1, joblib 1.4.2, json5 0.9.14, jsonpointer 2.4, jsonschema 4.21.1, jsonschema-specifications 2023.12.1, jupyter-archive 3.4.0, jupyter_client 7.4.9, jupyter_contrib_core 0.4.2, jupyter_contrib_nbextensions 0.7.0, jupyter_core 5.7.1, jupyter-events 0.9.0, jupyter-highlight-selected-word 0.2.0, jupyter-lsp 2.2.2, jupyter-nbextensions-configurator 0.6.3, jupyter_server 2.12.5, jupyter_server_terminals 0.5.2, jupyterlab 4.1.0, jupyterlab_pygments 0.3.0, jupyterlab_server 2.25.2, jupyterlab-widgets 3.0.9, keyring 23.5.0, launchpadlib 1.10.16, lazr.restfulclient 0.14.4, lazr.uri 1.0.6, locket 1.0.0, lxml 5.1.0, MarkupSafe 2.1.5, matplotlib-inline 0.1.6, mistune 3.0.2, more-itertools 8.10.0, mpmath 1.3.0, multidict 6.0.5, multiprocess 0.70.16, nbclassic 1.0.0, nbclient 0.9.0, nbconvert 7.14.2, nbformat 5.9.2, nest-asyncio 1.6.0, networkx 3.2.1, nltk 3.8.1, notebook 6.5.5, notebook_shim 0.2.3, numpy 1.26.3, nvidia-cublas-cu12 12.1.3.1, nvidia-cuda-cupti-cu12 12.1.105, nvidia-cuda-nvrtc-cu12 12.1.105, nvidia-cuda-runtime-cu12 12.1.105, nvidia-cudnn-cu12 8.9.2.26, nvidia-cufft-cu12 11.0.2.54, nvidia-curand-cu12 10.3.2.106, nvidia-cusolver-cu12 11.4.5.107, nvidia-cusparse-cu12 12.1.0.106, nvidia-nccl-cu12 2.19.3, nvidia-nvjitlink-cu12 12.3.101, nvidia-nvtx-cu12 12.1.105, oauthlib 3.2.0, overrides 7.7.0, packaging 23.2, pandas 2.2.2, pandocfilters 1.5.1, parso 0.8.3, partd 1.4.2, peft 0.11.1, pexpect 4.9.0, pillow 10.2.0, pip 24.1.2, platformdirs 4.2.0, pluggy 1.5.0, polars 1.1.0, prometheus-client 0.19.0, prompt-toolkit 3.0.43, protobuf 5.27.2, psutil 5.9.8, ptyprocess 0.7.0, pure-eval 0.2.2, pyarrow 16.1.0, pyarrow-hotfix 0.6, pyasn1 0.6.0, pycparser 2.21, Pygments 2.17.2, PyGObject 3.42.1, PyJWT 2.3.0, pyparsing 2.4.7, pytest 8.2.2, python-apt 2.4.0+ubuntu3, python-dateutil 2.8.2, python-json-logger 2.0.7, pytz 2024.1, PyYAML 6.0.1, pyzmq 24.0.1, referencing 0.33.0, regex 2024.5.15, requests 2.32.3, rfc3339-validator 0.1.4, rfc3986-validator 0.1.1, rpds-py 0.17.1, rsa 4.7.2, s3transfer 0.10.2, safetensors 0.4.3, scikit-learn 1.5.1, scipy 1.14.0, SecretStorage 3.3.1, Send2Trash 1.8.2, sentence-transformers 3.0.1, sentencepiece 0.2.0, setuptools 69.0.3, six 1.16.0, sniffio 1.3.0, soupsieve 2.5, stack-data 0.6.3, sympy 1.12, tabulate 0.9.0, terminado 0.18.0, threadpoolctl 3.5.0, tiktoken 0.7.0, tinycss2 1.2.1, tokenizers 0.19.1, tomli 2.0.1, toolz 0.12.1, torch 2.2.0, torchaudio 2.2.0, torchdata 0.7.1, torchtext 0.17.0, torchvision 0.17.0, tornado 6.4, tqdm 4.66.4, traitlets 5.14.1, transformers 4.42.4, triton 2.2.0, types-python-dateutil 2.8.19.20240106, typing_extensions 4.9.0, tzdata 2024.1, uri-template 1.3.0, urllib3 2.2.2, wadllib 1.3.6, wcwidth 0.2.13, webcolors 1.13, webencodings 0.5.1, websocket-client 1.7.0, wheel 0.42.0, widgetsnbextension 4.0.9, xxhash 3.4.1, yarl 1.9.4, zipp 1.0.0

Citation

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

BibTeX:

In Progress