IEEE MLSP

We are thrilled to announce that FCP Deputy Director, Prof Park Jihong, has been elected as one of the twelve new Technical Committee members for Machine Learning for Signal Processing (MLSP). This is a testament to Jihong's dedication and significant contributions to the MLSP community. We are confident that Jihong will continue to bring valuable insights and leadership to this field and the FCP team. Our sincere congratulations to Jihong! #SUTD #FCP #Congratulations #IEEE #MLSP #Leadership #Innovation

📢 The deadline for submitting an intention to bid for #MLSP2027 is approaching quickly: 27 November 2024 🔥 Looking forward to existing proposals!

Ready for #MLSP2025 🇹🇷🤩

I had the opportunity to present my research, "Integrating Visible Light for Enhanced Thermal Imaging: Parameterized Scene-Based Non-Uniformity Correction" at the IEEE International Workshop on Machine Learning for Signal Processing (MLSP) 2024.  This research, supported by Aselsan and conducted at Hacettepe University, focuses on improving thermal imaging using visible-light data, with real-time applications in autonomous driving and defense. #MLSP2024 #HacettepeUniversity #ASELSAN

Thrilled to announce that our paper on Ordinal Archetypal Analysis was recently published at IEEE MLSP 2024! A huge thank you to my fellow coauthors for the collaboration! More details on the work below! 👇

I was very proud to assist and encourage Quentin Langlois at the IEEE Workshop on Machine Learning for Signal Processing (IEEE MLSP 2024) at Imperium College in London. The event featured several inspiring keynotes and highly technical presentations. Quentin presented a paper on “Embedding for Motor Imagery Classification” during the poster session on the second day. This provided him with the opportunity to engage in fascinating discussions with several participants, including @Professor Saeid Sanei from Nottingham Trent University, @Shuang Ni from MILA Quebec AI Institute, @Teysir Baoueb and @Xiaoyu Bie from Telecom Paris, among others. #IEEE #MLSP2024 #MachineLearning #SignalProcessing #Research #Networking #AI #Innovation

I had the opportunity to present our work, 𝐑𝐄𝐒𝐓𝐀𝐃: 𝐑𝐄𝐜𝐨𝐧𝐬𝐭𝐫𝐮𝐜𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐒𝐢𝐦𝐢𝐥𝐚𝐫𝐢𝐭𝐲 𝐛𝐚𝐬𝐞𝐝 𝐓𝐫𝐚𝐧𝐬𝐟𝐨𝐫𝐦𝐞𝐫 𝐟𝐨𝐫 𝐓𝐢𝐦𝐞 𝐒𝐞𝐫𝐢𝐞𝐬 𝐀𝐧𝐨𝐦𝐚𝐥𝐲 𝐃𝐞𝐭𝐞𝐜𝐭𝐢𝐨𝐧, at the IEEE MLSP Conference held at Imperial College London! 🎤✨ Paper: https://lnkd.in/eG62QiaU Code: https://lnkd.in/e63zBggk It was an honor to share our findings with the Machine Learning and Signal Processing community, and engage in insightful discussions with fellow researchers and professionals. 🌍💡 I’m excited to apply the new ideas and insights gained from this experience to further my journey. 🚀 #AI #ML #AnomalyDetection #TimeSeries #RESTAD #Transformer #Research #IEEE #MLSP2024 #DataScience #MachineLearning

Had a chance to organize and lead the special session on "AI & Sustainability in Signal Processing" and co-chair the session in IEEE Conference on Machine Learning for Signal processing (MLSP) #MLSP2024 IEEE MLSP in Imperial College London on 23rd September 2024. Clealry, exploration of efficieny leads to contribute towards sustainable practices while using AI technology, as stated by 6 interesting papers from different fields (speech enhancement, audio captioning, speech deep fake, energy consumption measurement of AI models etc.) in our session: 1. ULTRA-LOW MEMORY SPEECH DENOISING USING QUANTIZATION-AWARE NEURAL ARCHITECTURE SEARCH by Anastasios Vafeiadis (NXP Semiconductors)*; Nick van de Waterlaat (NXP Semiconductors); Clément Castel (NXP Semiconductors); Bruno Defraene (NXP Semiconductors); Gerardo Daalderop (NXP Semiconductors); Sebastian Vogel (NXP Semiconductors); Christophe Macours (NXP Semiconductors) 2. SPEED: Scalable Preprocessing of EEG Data for Self-Supervised Learning Anders AGM Gjølbye (Technical University of Denmark)*; Lina Skerath (Technical University of Denmark); William T Lehn-Schiøler (Technical University of Denmark); Nicolas Langer (University of Zurich); Lars Kai Hansen (Technical University of Denmark) 3. Investigating Passive Filter Pruning for Efficient CNN-Transformer Audio Captioning Xuenan Xu (Shanghai Jiao Tong University)*; Arshdeep Singh (University of Surrey); Mengyue Wu (Shanghai Jiao Tong University); Wenwu Wang (University of Surrey); Mark D. Plumbley (University of Surrey) 4. Is that me? Using Speaker Identity to Detect Fake Speech Shilpa Chandra (IIT Mandi)*; Padmanabhan Rajan (IIT Mandi) 5. Normalizing energy consumption for hardware-independent evaluation Constance Douwes (Inria)*; Romain Serizel (Université de Lorraine) 6. Singular value decomposition for the efficient design of neural networks Vlad S Paul (University of Southampton)*; Philip Nelson (University of Southampton) Thanks to co-organizers Mark Plumbley Wenwu Wang Erick Sperandio and Constance Douwes for helping organizing the session smmothly. Thanks to all the presenters for sharing their research. University of Surrey Surrey Institute for People-Centred AI (PAI)Institute for Sustainability EPSRC Centre for Vision, Speech and Signal Processing at University of Surrey #AI #SUSTAINABILITY #AIforGood #research #conference #MLSP #Machinelearning #signalprocessing #EPSRC

On Friday, PAI was delighted to host insightful talks on audio and speech by four visiting researchers from France and India. Our guests were in London for the IEEE MLSP 2024 workshop. Three of our visiting speakers were from Télécom Paris Institut Polytechnique de Paris: XIAOYU BIE talked on: Learning Source Disentanglement in Neural Audio Codec: Discover how to improve audio compression and generation by separating different sound sources like speech and music! Teysir Baoueb taled on: WaveTransfer: A Flexible End-to-End Timbre Transfer with Diffusion: Learn about a new method for seamlessly changing the instrument sounds in music while keeping the melody intact. Xuanyu Zhuang talked on: Episodic fine-tuning prototypical networks for optimization-based few-shot learning: Explore a novel approach for training AI models to recognize sounds with very little data. Shilpa Chandra from the Indian Institute of Technology, Mandi spoke on the subject: Is that me? Using Speaker Identity to Detect Fake Speech: Dive into how speaker information can help identify deepfakes and other forms of synthetic speech. It was a happy exchange of ideas and knowledge, and we appreciated our visitors’ talks immensely. Thank you! #IEEE2024 #MLSP2024 Wenwu Wang | Mark Plumbley | Arshdeep Singh (PhD) | Gabriel Bibbó | Paul Bereuter | Yi Yuan