1- <!DOCTYPE html>
2- < html lang ="en ">
3- < head >
4- < meta charset ="UTF-8 ">
5- < meta name ="viewport " content ="width=device-width, initial-scale=1.0 ">
6- < title > Fundamental AI — BioIntelligence Lab</ title >
7-
8- < style >
9- body {
10- margin : 0 ;
11- font-family : "Inter" , sans-serif;
12- color : # 222 ;
13- line-height : 1.6 ;
14- background : # ffffff ;
15- }
16-
17- /* NAVIGATION BAR (same as homepage) */
18- nav {
19- width : 100% ;
20- padding : 20px 40px ;
21- box-sizing : border-box;
22- position : sticky;
23- top : 0 ;
24- background : rgba (255 , 255 , 255 , 0.90 );
25- backdrop-filter : blur (8px );
26- display : flex;
27- justify-content : space-between;
28- align-items : center;
29- border-bottom : 1px solid # eee ;
30- z-index : 1000 ;
31- }
32-
33- nav .logo img {
34- height : 45px ;
35- cursor : pointer;
36- }
37-
38- nav ul {
39- list-style : none;
40- display : flex;
41- gap : 30px ;
42- margin : 0 ;
43- padding : 0 ;
44- }
45-
46- nav ul li a {
47- text-decoration : none;
48- color : # 222 ;
49- font-weight : 500 ;
50- transition : color 0.2s ;
51- }
52-
53- nav ul li a : hover {
54- color : # 003d99 ;
55- }
56-
57- /* PAGE HEADER */
58- .header-block {
59- text-align : center;
60- padding : 120px 20px 80px ;
61- background : # f7f9fc ;
62- }
63-
64- .header-block h1 {
65- font-size : 2.6em ;
66- margin-bottom : 10px ;
67- font-weight : 700 ;
68- color : # 111 ;
69- }
70-
71- .header-block p {
72- font-size : 1.2em ;
73- color : # 555 ;
74- max-width : 800px ;
75- margin-left : auto;
76- margin-right : auto;
77- }
78-
79- /* SECTION TITLE */
80- h2 .section-title {
81- text-align : center;
82- margin-top : 70px ;
83- margin-bottom : 25px ;
84- font-size : 1.9em ;
85- font-weight : 600 ;
86- }
87-
88- /* CONTENT SECTIONS */
89- .content-block {
90- max-width : 900px ;
91- margin : 0 auto;
92- padding : 10px 20px 40px ;
93- color : # 333 ;
94- font-size : 1.05em ;
95- }
96-
97- .placeholder-img {
98- width : 100% ;
99- height : 220px ;
100- background : # e9eef5 ;
101- border-radius : 8px ;
102- margin : 30px 0 ;
103- display : flex;
104- align-items : center;
105- justify-content : center;
106- color : # 708090 ;
107- font-size : 1.1em ;
108- font-style : italic;
109- }
110-
111- /* PUBLICATION LIST */
112- .pub-section h3 {
113- margin-top : 40px ;
114- font-size : 1.3em ;
115- color : # 003d99 ;
116- }
117-
118- .pub-section ul {
119- margin-top : 10px ;
120- line-height : 1.55 ;
121- }
122- </ style >
123- </ head >
124-
125- < body >
126-
127- <!-- NAVIGATION BAR -->
128- < nav >
129- < div class ="logo ">
130- < a href ="index.html ">
131- < img src ="https://github.com/BioIntelligence-Lab/BioIntelligence-Lab.github.io/blob/main/images/Lab_logo3.png?raw=true " alt ="Lab Logo ">
132- </ a >
133- </ div >
134-
135- < ul >
136- < li > < a href ="index.html#research "> Research</ a > </ li >
137- < li > < a href ="index.html#tools "> Software</ a > </ li >
138- < li > < a href ="index.html#people "> People</ a > </ li >
139- < li > < a href ="index.html#contact "> Contact</ a > </ li >
140- </ ul >
141- </ nav >
142-
143- <!-- HEADER -->
144- < section class ="header-block ">
145- < h1 > Fundamental AI Research</ h1 >
146- < p >
147- Advancing the foundations of Artificial Intelligence through safety, trustworthiness, human–AI ecosystems,
148- and multi-agent autonomy. Our work pushes beyond application-driven AI to explore how intelligent systems
149- learn, collaborate, and evolve over time.
150- </ p >
151- </ section >
152-
153- <!-- AI SAFETY -->
154- < h2 class ="section-title "> AI Safety & Trustworthiness</ h2 >
155-
156- < div class ="content-block ">
157- < p >
158- This sub-area focuses on ensuring AI systems are reliable, transparent, and equitable.
159- We investigate algorithmic bias, uncertainty modeling, robustness to real-world variation,
160- security vulnerabilities, demographic leakage in foundation models, and safe use of generative AI
161- in clinical decision-making.
162- </ p >
163-
164- < div class ="placeholder-img "> [ Placeholder: Diagram on AI Safety / Bias / Robustness ]</ div >
165-
166- < div class ="pub-section ">
167- < h3 > Representative Publications</ h3 >
168- < ul >
169- < li > Beheshtian et al., Radiology, 2022 — Bias in pediatric bone age prediction.</ li >
170- < li > Bachina et al., Radiology, 2023 — Coarse race labels masking underdiagnosis patterns.</ li >
171- < li > Santomartino et al., Radiology: AI, 2024 — Stress testing and robustness evaluation.</ li >
172- < li > Trang et al., Emergency Radiology, 2024 — Sociodemographic bias in ICH detection.</ li >
173- < li > Kavandi et al., AJR, 2024 — Predictability of demographics from chest radiographs.</ li >
174- < li > Santomartino et al., Radiology, 2024 — Bias in NLP tools for radiology reports.</ li >
175- < li > Garin, Parekh, Sulam, Yi et al., Nature Medicine, 2023 — Need for demographic transparency.</ li >
176- < li > Yi et al., Radiology, 2025 — Best practices for evaluating algorithmic bias.</ li >
177- < li > Zheng, Jacobs, Parekh et al., arXiv, 2024 — Demographic predictability in CT embeddings.</ li >
178- < li > Zheng, Jacobs, Braverman, Parekh et al., arXiv, 2025 — Adversarial debiasing in CT models.</ li >
179- < li > Kulkarni et al., MIDL, 2024 — Hidden-in-plain-sight imperceptible bias attacks.</ li >
180- </ ul >
181- </ div >
182- </ div >
183-
184- <!-- HUMAN–AI ECOSYSTEM -->
185- < h2 class ="section-title "> Human–AI Ecosystem Modeling</ h2 >
186-
187- < div class ="content-block ">
188- < p >
189- We study how humans and AI systems can learn from each other, share experience, collaborate across institutions,
190- and form collective intelligence. This includes the development of SheLL (Shared Experience Lifelong Learning),
191- multi-agent reasoning frameworks, and the foundations needed to build autonomous research workflows.
192- </ p >
193-
194- < div class ="placeholder-img "> [ Placeholder: SheLL / Multi-Agent Collaboration Diagram ]</ div >
195-
196- < div class ="pub-section ">
197- < h3 > Representative Publications</ h3 >
198- < ul >
199- < li > Uwaeze, Kulkarni, Braverman, Jacobs, Parekh, ICCV 2025 — Counterfactual augmentation for equitable learning.</ li >
200- < li > Kulkarni et al., MIDL 2024 — Stealth bias attacks informing ecosystem resilience.</ li >
201- <!-- Add more as papers emerge -->
202- </ ul >
203- </ div >
204- </ div >
205-
206- </ body >
207- </ html >
1+ ---
2+ layout: default
3+ title: Fundamental AI
4+ ---
5+
6+ < section class ="header-block ">
7+ < h1 > Fundamental AI Research</ h1 >
8+ < p >
9+ Advancing the foundations of Artificial Intelligence through safety, trustworthiness, human–AI ecosystems, and
10+ multi-agent autonomy — exploring how intelligent systems learn, collaborate, and evolve over time.
11+ </ p >
12+ </ section >
13+
14+ <!-- AI SAFETY -->
15+ < h2 class ="section-title "> AI Safety & Trustworthiness</ h2 >
16+
17+ < div class ="content-block ">
18+ < p >
19+ We develop methods to ensure AI systems are reliable, transparent, and equitable — addressing bias, uncertainty,
20+ robustness to real-world clinical variation, security vulnerabilities, demographic leakage, and safe use of
21+ generative AI in radiology and beyond.
22+ </ p >
23+
24+ < div class ="placeholder-img "> [ Placeholder: Diagram — Bias / Robustness / Security ]</ div >
25+ </ div >
26+
27+ <!-- HUMAN-AI ECOSYSTEM -->
28+ < h2 class ="section-title "> Human–AI Ecosystem</ h2 >
29+
30+ < div class ="content-block ">
31+ < p >
32+ We investigate how AI systems can collaborate with each other and with humans — learning across sites, agents,
33+ and tasks. This includes the development of SheLL (Shared Experience Lifelong Learning), multi-agent reasoning,
34+ and foundations for autonomous research workflows.
35+ </ p >
36+
37+ < div class ="placeholder-img "> [ Placeholder: Diagram — SheLL / Multi-Agent Learning ]</ div >
38+ </ div >
0 commit comments