The rise of conversational AI and multimodal streaming applications has led to a significant demand for low-latency Text-to-Speech (TTS) systems. This work presents a multilingual low-latency model that leverages the functional decomposition principles of Kolmogorov-Arnold Networks (KANs) in modeling the nonlinear functions with lesser parameters and smaller computational graphs than multi-layer perceptrons (MLPs). To enhance the learning capabilities of the proposed compact model, we use a supervised auxiliary learning approach with multiple subtasks whose target labels are extracted from the speech signal. We use a multispeaker nonlinear vocoder to reconstruct the natural-sounding speech from the Melspectrograms. The proposed model achieved a very low latency with a real-time factor (RTF) of 0.0795 and a mean opinion score (MOS) of 4.09 for naturalness in the multilingual streaming TTS challenge organized as part of ICASSP-2025.

Two-factor authentication (TFA) provides an additional layer of protection to commonly-occurring password breaches. However, existing TFA methods, often involve special hardware interfaces, or require human effort which is prone to errors and acts as an adoption detractor for older adults and novice technology users. To address these limitations, we propose a zero-interaction, two-factor authentication (ZITA) protocol. In ZITA, the first factor is implemented using the conventional username and password methods. The second factor is completed without any human effort provided that the user is not accessing the service from an unregistered public device and a designated secondary device is physically co-present. To automate the second factor, ZITA exploits the long-term contact between the login device and the secondary device such as a smartphone. Moreover, to thwart man-in-the-middle and co-located attacks, ZITA incorporates a proximity verification test that relies on the randomness of ambient RF signals. Compared with other zero-effort TFA protocols, ZITA remains secure against advanced threats and does not require out-of-band sensors such as microphones, speakers, or photoplethysmography (PPG) sensors.

The use of Internet of Things (IoT) devices is higher than ever and is growing rapidly. Many IoT devices are manufactured by home appliance manufacturers where security and privacy is not the foremost concern. There does not exist a strict authentication method that verifies the identity of the device. This allows any rogue IoT device to authenticate and spoof various IoT device activities using compromised credentials. This paper addresses the issue by introducing a novel method for re- and continuous authentication utilizing a device-type classification as a new identity paradigm. We present RADTEC: a protocol for authenticating a device in a network by leveraging machine learning to classify the type of an IoT device attempting to connect to the network with an accuracy of over 95% in less than 0.65 milliseconds. We investigate multiple machine learning classifiers to infer the types of IoT devices and use them to develop a stricter and more efficient method for authentication.