Skip to content

Commit 1a365e4

Browse files
committed
--
1 parent 7427d7b commit 1a365e4

1 file changed

Lines changed: 35 additions & 35 deletions

File tree

nsf-iis-2348919/index.html

Lines changed: 35 additions & 35 deletions
Original file line numberDiff line numberDiff line change
@@ -82,10 +82,10 @@ <h5>Fair Data Structures</h5>
8282
<li><a href="#fair-count-min">Fair-Count-Min</a></li>
8383
<li><a href="#fair-epsilon-nets">Fair Epsilon Nets and Geometric Hitting Sets</a></li>
8484
<h5>Beyond Data Structures</h5>
85+
<li><a href="#unbiased-binning">Unbiased Binning</a></li>
8586
<li><a href="#fair-set-cover">Fair Set Cover</a></li>
8687
<li><a href="#dynamic-necklace-splitting">Dynamic Necklace Splitting</a></li>
8788
<li><a href="#mining-the-minoria">Mining the Minoria</a></li>
88-
<li><a href="#unbiased-binning">Unbiased Binning</a></li>
8989
</ol>
9090
<p class="toc-note">This list will be updated during the course of project.</p>
9191
</aside>
@@ -220,6 +220,40 @@ <h3>On Fair Epsilon Nets and Geometric Hitting Sets</h3>
220220

221221
<div class="pub-group-heading">Beyond Data Structures</div>
222222

223+
<article class="paper-card" id="unbiased-binning">
224+
<div class="paper-tag">VLDB 2026</div>
225+
<h3>Unbiased Binning</h3>
226+
<p class="paper-meta">
227+
Abolfazl Asudeh, Zeinab Asoodeh, Bita Asoodeh, and Omid Asudeh · <em>Proceedings of the VLDB Endowment</em>, Vol. 19, 2026
228+
</p>
229+
<div class="paper-feature">
230+
<figure class="paper-figure">
231+
<img src="imgs/UnbiasedBinning.jpg" alt="Illustration for the Unbiased Binning paper">
232+
</figure>
233+
234+
<div class="paper-feature-copy">
235+
<p>
236+
Discretizing raw features into bucketized attributes is a common step before sharing a dataset. However, this process can inadvertently introduce bias and amplify unfairness in downstream tasks.
237+
</p>
238+
<p>
239+
This work formulates the unbiased binning problem, which seeks bucketized attributes that satisfy group parity. The paper develops an efficient dynamic programming algorithm for equal-size binning, while also showing that perfect parity may come with a high price of fairness or may not exist when group distributions differ substantially.
240+
</p>
241+
<p>
242+
To support settings where small deviations from perfect parity are acceptable, the paper introduces <em>epsilon</em>-biased binning, which limits group disparities across buckets to at most <em>epsilon</em>. It first presents a quadratic-time dynamic programming algorithm, then addresses scalability with a local search strategy whose divide-and-conquer component finds valid solutions in near-linear time whenever one exists.
243+
</p>
244+
<p>
245+
The local search and divide-and-conquer algorithms are general and are not limited to equal-size binning. Extensive experiments on real-world and synthetic datasets confirm the efficiency of the algorithms and show that fairness-unaware binning can generate biased attribute representations, while fairness-aware binning can significantly reduce this bias with negligible price of fairness.
246+
</p>
247+
<p class="paper-citation">
248+
Citation: Abolfazl Asudeh, Zeinab Asoodeh, Bita Asoodeh, Omid Asudeh. <em>Unbiased Binning: Fairness-aware Attribute Representation</em>. Proceedings of the VLDB Endowment, Vol 19, 2026.
249+
</p>
250+
<p class="paper-links">
251+
<a href="https://arxiv.org/pdf/2509.21785" target="_blank" rel="noopener noreferrer">Technical Report</a>
252+
</p>
253+
</div>
254+
</div>
255+
</article>
256+
223257
<article class="paper-card" id="fair-set-cover">
224258
<div class="paper-tag">KDD 2025</div>
225259
<h3>Fair Set Cover</h3>
@@ -354,40 +388,6 @@ <h3>Mining the Minoria</h3>
354388
</div>
355389
</div>
356390
</article>
357-
358-
<article class="paper-card" id="unbiased-binning">
359-
<div class="paper-tag">VLDB 2026</div>
360-
<h3>Unbiased Binning</h3>
361-
<p class="paper-meta">
362-
Abolfazl Asudeh, Zeinab Asoodeh, Bita Asoodeh, and Omid Asudeh · <em>Proceedings of the VLDB Endowment</em>, Vol. 19, 2026
363-
</p>
364-
<div class="paper-feature">
365-
<figure class="paper-figure">
366-
<img src="imgs/UnbiasedBinning.jpg" alt="Illustration for the Unbiased Binning paper">
367-
</figure>
368-
369-
<div class="paper-feature-copy">
370-
<p>
371-
Discretizing raw features into bucketized attributes is a common step before sharing a dataset. However, this process can inadvertently introduce bias and amplify unfairness in downstream tasks.
372-
</p>
373-
<p>
374-
This work formulates the unbiased binning problem, which seeks bucketized attributes that satisfy group parity. The paper develops an efficient dynamic programming algorithm for equal-size binning, while also showing that perfect parity may come with a high price of fairness or may not exist when group distributions differ substantially.
375-
</p>
376-
<p>
377-
To support settings where small deviations from perfect parity are acceptable, the paper introduces <em>epsilon</em>-biased binning, which limits group disparities across buckets to at most <em>epsilon</em>. It first presents a quadratic-time dynamic programming algorithm, then addresses scalability with a local search strategy whose divide-and-conquer component finds valid solutions in near-linear time whenever one exists.
378-
</p>
379-
<p>
380-
The local search and divide-and-conquer algorithms are general and are not limited to equal-size binning. Extensive experiments on real-world and synthetic datasets confirm the efficiency of the algorithms and show that fairness-unaware binning can generate biased attribute representations, while fairness-aware binning can significantly reduce this bias with negligible price of fairness.
381-
</p>
382-
<p class="paper-citation">
383-
Citation: Abolfazl Asudeh, Zeinab Asoodeh, Bita Asoodeh, Omid Asudeh. <em>Unbiased Binning: Fairness-aware Attribute Representation</em>. Proceedings of the VLDB Endowment, Vol 19, 2026.
384-
</p>
385-
<p class="paper-links">
386-
<a href="https://arxiv.org/pdf/2509.21785" target="_blank" rel="noopener noreferrer">Technical Report</a>
387-
</p>
388-
</div>
389-
</div>
390-
</article>
391391
</div>
392392
</div>
393393
</div>

0 commit comments

Comments
 (0)