Papers › Robust Concept Erasure via Kernelized Rate-Distortion Maximization

Robust Concept Erasure via Kernelized Rate-Distortion Maximization

30 Nov 2023NeurIPS 2023 11arXiv:2312.00194archive 2025-07-28

Somnath Basu Roy Chowdhury, Nicholas Monath, Avinava Dubey, Amr Ahmed, Snigdha Chaturvedi

Distributed representations provide a vector space that captures meaningful relationships between data instances. The distributed nature of these representations, however, entangles together multiple attributes or concepts of data instances (e.g., the topic or sentiment of a text, characteristics of the author (age, gender, etc), etc). Recent work has proposed the task of concept erasure, in which rather than making a concept predictable, the goal is to remove an attribute from distributed representations while retaining other information from the original representation space as much as possible. In this paper, we propose a new distance metric learning-based objective, the Kernelized Rate-Distortion Maximizer (KRaM), for performing concept erasure. KRaM fits a transformation of representations to match a specified distance measure (defined by a labeled concept to erase) using a modified rate-distortion function. Specifically, KRaM's objective function aims to make instances with similar concept labels dissimilar in the learned representation space while retaining other information. We find that optimizing KRaM effectively erases various types of concepts: categorical, continuous, and vector-valued variables from data representations across diverse domains. We also provide a theoretical analysis of several properties of KRaM's objective. To assess the quality of the learned representations, we propose an alignment score to evaluate their similarity with the original representation space. Additionally, we conduct experiments to showcase KRaM's efficacy in various settings, from erasing binary gender variables in word embeddings to vector-valued variables in GPT-3 representations.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 21 code samples harvested from 1 repository linked to this paper: 20 ran (of which 0 constructed an object rather than computing a result; 15 with no instrument failure: 4 honoured, 0 violated, 11 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified. Where Syntology's instrument failed: 5 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 21 samples from 1 repository, 20 ran (15 with no instrument failure; 5 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

brcsomnath/kram officialmentioned in paperpytorchMITSyntology: 21 samples, 20 ran (5 instrument-failed) · 1 unverified 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

21 samples harvested: 20 ran (of which 0 constructed an object rather than computing a result; 15 with no instrument failure: 4 honoured, 0 violated, 11 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

4ran · honoured contract
5ran · our draft was wrong
11ran
1unverified

Licence: 2 of the 21 samples are pointer only, meaning Syntology does not serve that copy's text. For 2 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from brcsomnath/kram. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

brcsomnath/kram official repository 21 samples, 20 ran (5 instrument-failed) · 1 unverified
RateDistortion src/loss.py ran MIT (permissive) · 49a93e98ae41a86f · report fixture: signature W:Tensor/2/float, device:str/0/str · returned Tensor · level recorded 2026-09-24
RateDistortionContinuous src/loss.py ran MIT (permissive) · efd704682a41f0a3 · report fixture: signature X:Tensor/2/float, X_raw:Tensor/2/float, kernel:Tensor/2/float, device:str/0/str, scale:float/0/float · returned tuple · level recorded 2026-09-24
cauchy_kernel src/metrics.py ran MIT (permissive) · 8e8e1bf9b627b971 · report fixture: signature tensor:Tensor/2/float, sigma:float/0/float, data_type:str/0/str · returned Tensor · level recorded 2026-09-24
classify in 27 papers src/utils.py ran MIT (permissive) · 590b4d2c85eae0a0 · report fixture: signature x_train:ndarray/2/float, y_train:ndarray/1/int, x_test:ndarray/2/float, y_test:ndarray/1/int, seed:int/0/int · returned float · level recorded 2026-09-24
correlation in 16 papers src/evaluate.py ran MIT (permissive) · 8a14ef4db6394aa7 · report fixture: signature Z_before:ndarray/2/float, Z_after:ndarray/2/float, k:int/0/int · returned list · level recorded 2026-09-24
debias_by_specific_directions in 7 papers src/baselines/inlp.py ran · our draft was wrong fingerprinted MIT (permissive) · 8913e4e11f92191f · report fixture: signature directions:list/1/ndarray, input_dim:int/0/int · returned ndarray · level recorded 2026-09-24
evaluate_regression src/utils.py ran MIT (permissive) · b5e6d7d6e3f5eb54 · report fixture: signature x_train:ndarray/2/float, z_train:ndarray/1/float, x_test:ndarray/2/float, z_test:ndarray/1/float, seed:int/0/int · returned tuple · level recorded 2026-09-24
gaussian_kernel in 51 papers
census: 9 implementations from 9 papers, 9 distinct outputs across 6 buckets, one shared input per bucket
src/metrics.py ran MIT (permissive) · c4b2effb5d8110a1 · report fixture: signature tensor:Tensor/2/float, sigma:float/0/float, data_type:str/0/str · returned Tensor · level recorded 2026-09-24
get_projection_to_intersection_of_nullspaces in 7 papers src/baselines/inlp.py ran · our draft was wrong fingerprinted MIT (permissive) · b26df1b3be7a084e · report fixture: signature rowspace_projection_matrices:list/1/ndarray, input_dim:int/0/int · returned ndarray · level recorded 2026-09-24
get_representations_from_loader src/baselines/farm.py ran · honoured contract MIT (permissive) · 3f72ad49af8d0e6b · report fixture: signature loader:list/1/tuple · returned ndarray · level recorded 2026-09-24
get_representations_from_loader src/baselines/mi_debiasing.py ran · honoured contract MIT (permissive) · 4484e72f4b964121 · report fixture: signature loader:list/1/tuple · returned ndarray · level recorded 2026-09-24
get_rowspace_projection in 7 papers src/baselines/inlp.py ran · honoured contract fingerprinted MIT (permissive) · f75e44f7fc419025 · report fixture: signature W:ndarray/2/float · returned ndarray · level recorded 2026-09-24
knn in 87 papers
census: 12 implementations from 17 papers, 6 distinct outputs across 4 buckets, one shared input per bucket
src/evaluate.py ran · our draft was wrong fingerprinted MIT (permissive) · 0ff8884ceed2cb83 · report fixture: signature Z:ndarray/2/float, k:int/0/int · returned ndarray · level recorded 2026-09-24
label_to_membership src/loss.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · abef5bdd6d4d119a · report fixture: signature targets:ndarray/1/int, num_classes:int/0/int · returned ndarray · level recorded 2026-09-22
laplace_kernel in 5 papers src/metrics.py ran MIT (permissive) · 4b0b366504cec402 · report fixture: signature tensor:Tensor/1/float, sigma:float/0/float, data_type:str/0/str · returned Tensor · level recorded 2026-09-24
load_data in 700 papers src/data_load.py ran MIT (permissive) · 69fd31f5a04ad9e2 · report fixture: signature path:str/0/str, size:int/0/int, ratio:float/0/float · returned tuple · level recorded 2026-09-24
load_deepmoji src/data_load.py ran MIT (permissive) · 55c2f56cc152566e · report fixture: signature path:str/0/str, ratio:float/0/float · returned tuple · level recorded 2026-09-24
one_hot in 150 papers
census: 18 implementations from 25 papers, 10 distinct outputs across 6 buckets, one shared input per bucket
src/loss.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · 62dfecd92146a155 · report fixture: signature labels_int:list/1/int, n_classes:int/0/int · returned Tensor · level recorded 2026-09-22
pairwise_dist in 10 papers src/baselines/farm.py ran · honoured contract fingerprinted MIT (permissive) · c27d027c708f2a99 · report fixture: signature Z:ndarray/2/float · returned ndarray · level recorded 2026-09-24
regression_predict src/utils.py ran MIT (permissive) · e6ca500737cb0b1c · report fixture: signature x_train:ndarray/2/float, z_train:ndarray/2/float, x_test:ndarray/2/float, z_test:ndarray/2/float, seed:int/0/int · returned ndarray · level recorded 2026-09-24
read_dataset in 32 papers src/data_load.py unverified MIT (permissive) · d127dafd8f5d0f8a · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for Kernelized Rate-Distortion Maximizer (KRaM) (graph method: Kernelized Rate-Distortion Maximizer (KRaM)). 2 bricks; individually on a fixture: 2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.

  1. RateDistortionContinuous entry · level 1 · ran · efd704682a41f0a3
  2. RateDistortion helper · level 1 · ran · 49a93e98ae41a86f

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2312.00194") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

AttributeMetric LearningWord Embeddings

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamAttentionAttention DropoutBPECosine AnnealingDense ConnectionsDropoutGPT-3Layer NormalizationLinear LayerLinear Warmup With Cosine AnnealingMulti-Head AttentionResidual ConnectionSoftmaxWeight Decay

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