a conversation with Binghamton faculty Adnan Siraj Rakin assistant professor in the School of Computing. At 12:15-1:15 p.m. Friday, Nov.7, in AD 148. Rakin will present backslash “Resolving the tension between model efficiency and adversarial robustness for AI algorithms. backslash “ AdderNet emerges with superior hardware efficiencies compared to Convolutional Neural Networks ( CNNs ) . Nevertheless, AdderNet backslash ’s security exploration remains underexplored. To utilize the hardware efficiency of AdderNet and at the same time improve resiliency against Bit-Flip Attack ( BFA ) , we propose a novel Non-Negative AdderNet ( NNAN ) model. Our proposed NNAN makes BFA obsolete and requires many more attack cycles to be compromised against BFA plus , successfully enabling the design of both hardware-efficient and secure AdderNet against BFA. Banner for Data Science Transdisciplinary Area of Excellence presents DS-TAE’s Data Salon: Friday, Nov. 7 featuring Binghamton faculty Adnan Siraj Rakin

Data Science Transdisciplinary Area of Excellence presents DS-TAE's Data Salon: Friday, Nov. 7 featuring Binghamton faculty Adnan Siraj Rakin

by Transdisciplinary Areas of Excellence

Speaker / Lecture Academic BU Scholars*- High-Impact Learning BU Scholars*- Intellectual Curio... Career Exploration Diversity/Cultural Political Professional Development Research

Fri, Nov 7, 2025

11:55 AM – 1:30 PM EST (GMT-5)

Add to Calendar
Friday, Nov. 7, 2025, 12:15-1:15 pm in AD 148 Couper Administration building

Private Location (sign in to display)

0
Registered

Registration

Details

Data Science Transdisciplinary Area of Excellence presents DS-TAE backslash ’s Data Salon: Nov. 7 The Data Science Transdisciplinary Area of Excellence ( DS-TAE ) presents its third installment of Data Salon, a conversation with Binghamton faculty Adnan Siraj Rakin assistant professor in the School of Computing. At 12:15-1:15 p.m. Friday, Nov.7, in AD 148. Rakin will present backslash “Resolving the tension between model efficiency and adversarial robustness for AI algorithms. backslash “ AdderNet emerges with superior hardware efficiencies compared to Convolutional Neural Networks ( CNNs ) . Nevertheless, AdderNet backslash ’s security exploration remains underexplored. To utilize the hardware efficiency of AdderNet and at the same time improve resiliency against Bit-Flip Attack ( BFA ) , we propose a novel Non-Negative AdderNet ( NNAN ) model. Our proposed NNAN makes BFA obsolete and requires many more attack cycles to be compromised against BFA plus , successfully enabling the design of both hardware-efficient and secure AdderNet against BFA. The Data salon session features a short talk to spark ideas, followed by generous time for open discussion and social interaction. The emphasis is not on presentations, but on creating space for dialogue, connection, and the possibility of new collaborations. Forward this announcement to new faculty members who may be interested in joining the Data Science TAE listserv and becoming part of the community. Lunch is provided for the participants. In-person participation is encouraged to make the most of the conversations and connections. However, a Zoom option will also be available here: https: slash slash binghamton.zoom.us slash j slash 93495116544. Flyer for Data Science Transdisciplinary Area of Excellence presents DS-TAE’s Data Salon: Friday, Nov. 7 featuring Binghamton faculty Adnan Siraj Rakin
The Data Science Transdisciplinary Area of Excellence (DS-TAE) presents its third installment of Data Salon, a conversation with Binghamton faculty Adnan Siraj Rakin assistant professor in the School of Computing. At 12:15-1:15 p.m. Friday, Nov.7, in AD 148. Rakin will present "Resolving the tension between model efficiency and adversarial robustness for AI algorithms." AdderNet emerges with superior hardware efficiencies compared to Convolutional Neural Networks (CNNs). Nevertheless, AdderNet's security exploration remains underexplored. To utilize the hardware efficiency of AdderNet and at the same time improve resiliency against Bit-Flip Attack (BFA), we propose a novel Non-Negative AdderNet (NNAN) model. Our proposed NNAN makes BFA obsolete and requires many more attack cycles to be compromised against BFA+, successfully enabling the design of both hardware-efficient and secure AdderNet against BFA.

The Data salon session features a short talk to spark ideas, followed by generous time for open discussion and social interaction. The emphasis is not on presentations, but on creating space for dialogue, connection, and the possibility of new collaborations. Forward this announcement to new faculty members who may be interested in joining the Data Science TAE listserv and becoming part of the community.

Lunch is provided for the participants. In-person participation is encouraged to make the most of the conversations and connections. However, a Zoom option will also be available here: https://binghamton.zoom.us/j/93495116544

File Attachments: data_salon_Adnan_Rakin_Nov_7_2025

Food Provided (light lunch and refreshments)

Hosted By

Transdisciplinary Areas of Excellence | Website | View More Events

Xingye Qiao

Contact the organizers