-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathblog.html
More file actions
273 lines (213 loc) · 15.4 KB
/
blog.html
File metadata and controls
273 lines (213 loc) · 15.4 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="description" content="Homepage of KleinLab">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>KleinLab - Homepage</title>
<link href="./libs/bootstrap-3.4.1/css/bootstrap.css" rel="stylesheet">
<link href="./libs/fontawesome-6.2.0/css/all.min.css" rel="stylesheet">
<link href="./style.css" rel="stylesheet">
<link href='https://fonts.googleapis.com/css?family=Didact Gothic' rel='stylesheet'>
<style>
body {
font-family: 'Didact Gothic';
}
</style>
<link rel="icon" href="images/logo_light_wo_frame.png">
</head>
<script async src="https://www.googletagmanager.com/gtag/js?id=G-E2B3ZSQM8N"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'G-E2B3ZSQM8N');
</script>
<body data-spy="scroll" data-target="#myScrollspy" data-offset="80">
<a href="#" id="scrollbutton" class="hidden-md hidden-lg hidden-xl" style="display: none;"><span></span></a>
<div class="container">
<div style="margin-top: 15px; margin-bottom: 15px;">
<img class="filter-img" src="images/lettering_dark.png" alt="Group logo" style="height: 40px;">
<h2 style="color:#4c8a90; display: inline; vertical-align:middle; margin-left: 10px; "> - Methods for Big Data</h2>
</div>
<div id="navbar"></div>
<script>
fetch('/navbar.html')
.then(response => response.text())
.then(data => {
document.getElementById('navbar').innerHTML = data;
// Highlight current page
const path = window.location.pathname;
const navLinks = document.querySelectorAll('.nav-link');
navLinks.forEach(link => {
const href = link.getAttribute('href');
if (href === path || (path === '/' && href === '/index.html')) {
link.classList.add('active');
}
});
// Highlight parent for subpages
if (path.startsWith('/subpages_pubs/')) {
document.querySelector('a[href="/publications.html"]').classList.add('active');
}
if (path.startsWith('/subpages_research/')) {
document.querySelector('a[href="/research.html"]').classList.add('active');
}
});
</script>
<!-- ------------------------------------------------------------------------------------------------------------------------------------ -->
<!-- ------------------------------------------------------------------------------------------------------------------------------------ -->
<!-- ------------------------------------------------------------------------------------------------------------------------------------ -->
<!-- ------------------------------------------------------------------------------------------------------------------------------------ -->
<!-- Blog -->
<section id="Blog" style="font-size:medium;">
<h2>Blog <span class="fa fa-bookmark" aria-hidden="true" style="float:left;padding-right:15px;padding-top:6px;font-size: 75%;"></span></h2>
<ul class="fa-ul list-group" style="margin-left:15px; margin-right:15px">
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/DependentCensoring/figures/figure_1.png" style="width: 90%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/DependentCensoring/post.html"><h4>Modeling Dependent Censoring in Time-to-Event Data using Boosting Copula Regression </h4></a> (By Annika Strömer and Nadja Klein, posted on October 24, 2025)
<p style="text-align: justify; margin-top: 10px;">This paper presents a boosting copula regression approach to model dependent censoring in survival analysis. By jointly modeling survival and censoring times together and including covariate effects, our approach can flexibly capture and adjust for dependencies between them. This allows for a more realistic understanding of complex biomedical processes, even in high-dimensional data situations.</p>
</div>
</div>
</li>
<br>
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/BoostDAGs/figures/two_d_sem.png" style="width: 75%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/BoostDAGs/post.html"><h4>Boosting Causal Additive Models </h4></a> (By Maximilian Kertel and Nadja Klein, posted on September 13, 2025)
<p style="text-align: justify; margin-top: 10px;">This paper introduces a novel method for causal discovery from purely observational data. The research focuses on a class of models known as Causal Additive Models (CAMs), which assume that variables follow a Structural Equation Model (SEM) with additive noise. We present a boosting-based approach that can consistently identify the causal ordering of variables.</p>
</div>
</div>
</li>
<br>
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/HetSNGP/figures/arch_shift.drawio.png" style="width: 90%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/HetSNGP/post.html"><h4>Uncertainty-Aware Trajectory Prediction via Rule-Regularized Heteroscedastic Deep Classification </h4></a> (By Christian Schlauch and Nadja Klein, posted on July 18, 2025)
<p style="text-align: justify; margin-top: 10px;">We propose SHIFT (Spectral Heteroscedastic Informed Forecasting for Trajectories), a novel training framework for motion prediction in autonomous driving. SHIFT combines well-calibrated uncertainty modeling through last-layer heteroscedastic Gaussian processes and informative priors derived from automated rule extraction.</p>
</div>
</div>
</li>
<br>
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/PosteriorConsistencyPsplines/figures/PriorDraws.png" style="width: 90%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/PosteriorConsistencyPsplines/post.html"><h4>Posterior Concentration Rates for Bayesian Penalized Splines</h4></a> (By Paul Bach and Nadja Klein, posted on May 05, 2025)
<p style="text-align: justify; margin-top: 10px;">We close a gap in the literature and study the asymptotic properties of Bayesian penalized splines. We show that near-optimal posterior concentration can be achieved if the order of the penalty matches the regularity of the unknown function and if the hyperprior on the smoothing variance is appropriately chosen.</p>
</div>
</div>
</li>
<br>
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/TrulyMultDistReg/Figures/traffic_timepoints.png" style="width: 90%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/TrulyMultDistReg/post.html"><h4>Truly Multivariate Structured Additive Distributional Regression</h4></a> (By Lucas Kock and Nadja Klein, posted on April 11, 2025)
<p style="text-align: justify; margin-top: 10px;">We propose a truly multivariate distributional regression model. Building on copula regression, we model the dependence structure of the response through a Gaussian copula, while the marginal distributions can vary across components. Each parameter of the flexible distribution is linked to covariates. Our model is highly parameterized but estimation becomes feasible with Bayesian inference employing shrinkage priors.</p>
</div>
</div>
</li>
<br>
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/BoostDistrCopReg/figures/Fig1.png" style="width: 90%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/BoostDistrCopReg/post.html"><h4>Boosting distributional copula regression for bivariate binary, discrete and mixed responses</h4></a> (By Guillermo Briseño Sanchez and Nadja Klein, posted on April 01, 2025)
<p style="text-align: justify; margin-top: 10px;">We extend the current component-wise boosting methodology for distributional copula regression to accommodate prevalent type of variables in biomedical research: bivariate binary, bivariate discrete, as well as mixed binary-continuous. One of the benefits of our boosting approach is its ability to handle high-dimensional input variables. We illustrate our proposed method by analysing three challenging datasets as well as a number of experiments.</p>
</div>
</div>
</li>
<br>
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/SemiSupervisedOD/Figures/Fig1.png" style="width: 90%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/SemiSupervisedOD/post.html"><h4>Building Blocks for Robust and Effective Semi-Supervised Real-World Object Detection</h4></a> (By Moussa Kassem Sbeyti and Nadja Klein, posted on March 27, 2025)
<p style="text-align: justify; margin-top: 10px;">Semi-supervised object detection (SSOD) reduces the need for extensive labeled datasets but faces real-world challenges such as class imbalance, label noise, and unreliable pseudo-labels. We investigate the impact of these challenges on SSOD and propose four novel data-centric building blocks—Rare Class Collage (RCC), Rare Class Focus (RCF), Ground Truth Label Correction (GLC), and Pseudo-Label Selection (PLS)—that improve SSOD frameworks. Our methods achieve up to a 6% performance increase by enhancing label quality and class representation while maintaining computational efficiency.</p>
</div>
</div>
</li>
<br>
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/PriorsMotion/figures/Fig3.png" style="width: 90%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/PriorsMotion/post.html"><h4>Leveraging Informative Priors for Robust and Data-Efficient Motion Prediction</h4></a> (by Christian Schlauch and Nadja Klein, posted on March 13, 2025)
<p style="text-align: justify; margin-top: 10px;">Together with the Continental Automotive GmbH, we developed an informed learning approach for deep learning motion prediction models in autonomous driving. Using probabilistic models to learn informative priors from synthetic training labels, we regularize the training on observed data, improving data-efficiency and robustness of state-of-the-art models.</p>
</div>
</div>
</li>
<br>
<li class="list-group-item">
<div class="row" style="border: 1px solid black;">
<div class="column2" style="width: 30%; margin-left: 10px; margin-right: 10px; margin-top: 10px; margin-bottom: 10px; display: flex; align-items: center; justify-content: center;">
<img src="blog_posts/WeedsGalore/overview.png" style="width: 90%; display: block; margin-left: auto; margin-right: auto;">
</div>
<div class="column2" style="width: 70%; margin-right: 10px; margin-top: 10px; margin-bottom: 10px;">
<a href="blog_posts/WeedsGalore/post.html"><h4>A UAV-based dataset for plant monitoring in maize fields</h4></a> (by Ekin Celikkan and Nadja Klein, posted on February 21, 2025)
<p style="text-align: justify; margin-top: 10px;"> WeedsGalore is a novel agricultural dataset, focusing on crop and weed segmentation from UAV imagery in croplands. The high plant cover (in terms of both number and area) and the availability of five spectral bands are key distinguishing characters of this dataset. When used with state-of-the-art deep neural networks, it leads to models that can effectively segment weeds, even in unseen fields.</p>
</div>
</div>
</li>
<br>
</ul>
</section>
<!-- Imprint and data protection -->
<div style="display: inline-block; text-align: right; width: 100%; margin-top: 15px;">
<a href="https://www.scc.kit.edu/en/legals.php">Imprint</a> / <a href="https://www.scc.kit.edu/en/datenschutz.php">Privacy Policy</a>
</div>
</div>
</div>
</div>
</div>
<!-- jquery -->
<script src="./libs/jquery-3.6.1.min.js"></script>
<!-- boostrap -->
<script src="./libs/bootstrap-3.4.1/js/bootstrap.min.js"></script>
<!-- Functionality for popover of buttons -->
<script>
$(document).ready(function(){
$('[data-toggle="popover"]').popover();
});
</script>
<!-- Functionality for scroll-up button -->
<script>
$(document).ready(function(){
if ($(this).scrollTop() > 100) {
$('#scrollbutton').fadeIn();
}
$(window).scroll(function(){
if ($(this).scrollTop() > 100) {
$('#scrollbutton').fadeIn();
} else {
$('#scrollbutton').fadeOut();
}
});
$('#scrollbutton').click(function(){
$("html, body").animate({ scrollTop: 0 }, 300);
return false;
});
});
</script>
</body>
</html>