CNERG - Complex Networks Research Group, IIT Kharagpur

Home
India
Kharagpur
CNERG - Complex Networks Research Group, IIT Kharagpur

Complex Networks Research Group (CNeRG), a research lab of CSE, IIT KGP, interested in understanding Follow the page for our updates.

We computer scientists, with a long legacy in graph theory and algorithms, strongly feel that we have a lot to contribute to and learn from the very recent and extremely potent field of Complex Networks. This is the primary reason for the inception of the Complex Networks Research Group (CNeRG) at the Department of Computer Science and Engineering, IIT Kharagpur, India. Complex network research gr

oup is specially interested in understanding the structure and dynamics of large networked system with focus on online social systems. Various techniques from network science, graph theory, natural language processing, machine learning, communication system engineering, data analytics are leveraged to solve hard and interesting problems in this area. Specifically CNeRG presently works on:

(a) OSN like Twitter, Reddit etc. (b) Citation Network

(c) Socio-Mobile Application

Some of the topics which we are currently pursuing includes community detection, information diffusion, opinion dynamics, social search and recommendation.

07/08/2025

Like always, Indoml 2025 is back again with its spectacular array of speakers!

27/06/2025

This year CNeRG had a very solid presence at icwsm. Naquee presented his work on zero-shot VLMs performance in hate meme detection, Somnath preseted his paper on jailbreaking LLMs through instruction-centric prompts and Siddharth presented his work on disparity-accuracy trade-offs in face recognition systems. Animesh Mukherjee chaired two sessions on Politics and Social Media and Online Polarization. Overall it was a great fun and learning experience at Copenhagen.

27/03/2025

Paramita Das defended her thesis successfully. Prof Srinath Srinivasa, IIIT Bangalore conducted the defense and highly appreciated the thesis, especially her consolidated efforts to monitor and enable the well-being of the Wikipedia ecosystem. CNeRG wishes her all the best for her future endeavours.

Supervisor's note from Animesh Mukherjee
Paramita completes my baker's dozen. This marks the culmination of years of hard work, perseverance, and intellectual growth, and I could not be prouder of how far she has come.

Her journey has been one of immense dedication and curiosity, and it has been a privilege to witness her transformation into an independent researcher. The rigor, depth, and enthusiasm she brought to her research on Wikipedia will undoubtedly leave a lasting impact. She will undoubtedly continue to achieve great things in the next phase of her career.

My sincerest wish is that as you move forward, carry with you the same passion and determination that defined your time here. Wishing you all the success and fulfilment in your future endeavours -- this is just the beginning! And of course, I will forever owe you a treat with "dimer jhol bhaat" ... stay blessed.

01/03/2025

Somnath Banerjee at AAAI - Association for the Advancement of Artificial Intelligence 2025. He is there to present his work: SafeInfer: Context Adaptive Decoding Time Safety Alignment for Large Language Models (https://arxiv.org/abs/2406.12274). This work is a collaboration with Microsoft IDC within the MAPG scheme. Somnath recounts that the fun part was to explain his work to THE MAN with him in the frame.

07/02/2025

IIT Kharagpur invites PhD applications for the Autumn 2025-2026: https://www.iitkgp.ac.in/phd-admission
Apart from the regular PhD programme, IIT KGP also offers a programme for Working Professionals (more information here: https://www.iitkgp.ac.in/assets/pdf/acad_phd_wpg.pdf)

At CNeRG we promote an absolutely brilliant research culture always motivating the scholars to publish in Core A*/A conferences that define the Computer Science clout. In addition, we organize weekly reading groups where a pair of members present recently published papers at top conferences for the benefit of the whole group. Finally, we organise a yearly retreat where we hide from the regular hustle-bustle and introspect on our yearly activities to prepare better for the future. We maintain a tremendously high work ethic and a friendly environment -- scholars work "with" us as collaborators and not "under" us. We cordially invite you to apply for the programme and join us at CNeRG.

03/02/2025

Earlier today Punyajoy Saha defended his PhD thesis. Prof Vikram Goyal from IIIT Delhi conducted the defense. He was supervised by Animesh Mukherjee
He worked on developing frameworks for effective content moderation on social media platforms, with a focus on detecting and mitigating harmful speech. His research examined the prevalence of fear speech in Indian WhatsApp groups and compared it to hate speech on Gab. He addressed the challenges in data annotation, particularly the impact of message length and complex language on labeling accuracy. His findings revealed that fear speech generates higher engagement than hate speech and demonstrated the effectiveness of counterspeech in reducing harmful content. Additionally, his study underscored the critical role of explainability in AI models, highlighting how improved transparency enhances both model performance and user trust in content moderation systems.

Supervisor's note.

Punyajoy completes my dozen! As an academic, there are few joys greater than guiding bright and passionate young minds. However, getting educated by one's student is a different kind of joy altogether. Yes, Punyajoy was more a companion than a student - we fiercely debated while formulating problems, wrote papers together correcting each other, rejoiced acceptances and (mourned?) rejections. With him, I had a remarkable collaboration constantly publishing in all "big" venues. While people know me for the impactful fear speech reasearch, it was actually his brainchild and all credit goes to him.

Beyond his academic achievements, what truly set Punyajoy apart was his warmth, humility, and collaborative spirit. His enthusiasm was infectious, his determination inspiring, and his sense of humor a constant source of joy in the lab.

As he steps into the next phase of his career, I have no doubt that he will continue to make remarkable contributions. Punyajoy, your journey is just beginning, and the world awaits your brilliance. I am immensely proud of you and excited to see where your path leads. Wishing you success, fulfilment, and happiness always!

24/01/2025

Two Project Scientist positions are available for research on Responsible AI. Please apply. Link to apply: https://erp.iitkgp.ac.in/SRICStaffRecruitment/Login.jsp?ref_no=IIT/SRIC/R/TIE/2025/102

31/12/2024

Heartiest Congratulations to Sagnik Basu for receiving the Departmental Silver Medal at the 70th Convocation. He has topped the outgoing MTech batch. He did his MTech project with Animesh Mukherjee on studying biases in open-source face recognition systems. He has now registered for the joint MTech-PhD programme of the Institute and is working on the safety of visual language models.

25/12/2024

This year Indoml was held at BITS Pilani Goa campus. CNeRG members participated in various important roles.

Niloy Ganguly was the General Chair.
Pawan Goyal was the PAN India Forum Chair.
Animesh Mukherjee was the Sponsorship Chair.

Multiple CNeRG Alumni were also involved in the programme:
Surjya Ghosh was the Graduate Forum Chair.
Abhijnan Chakraborty and Mayank Singh were the Early Career Forum Chairs.
Soumi Das and Rajdeep Mukherjee were the Datathon Chairs.
Abir De was the Sponsorship Chair.
Rima Hazra was the Web Chair.

We had an awesome three day event packed with talks from stellar speakers in the area of AI/ML, graduate forum, pan India forum, early career forum, mentoring sessions and the datathon.

The event was attended by 200+ students across India whose travel, accommodation and all other logistics were fully sponsored by Indoml which has been the flagship mode of operation since its inception -- a model rarely noticed in the Indian academia. This year the conference spent nearly USD 25K (i.e., nearly half of the budget) to extend this facility to the students across India. The participation from the students was overwhelming and whole hearted since there were no strings attached -- come, listen, interact and through this process pave your way for higher studies. We strongly wish to see a steady growth in this culture over the years.

30/11/2024

Kishalay Das has been recognized as one of the top reviewers at the Learning on Graphs Conference 2024.

28/11/2024

Synopsis seminar

Sugandh Pargal gave her Synopsis seminar titled "Driving Behavior Under Diverse Micro-cultures: Developing Ubiquitous Systems for Safe Driving" on 27th November 2024.

Summary of her thesis : Road safety is a pressing global concern, with millions of lives lost annually due to road accidents caused by unsafe driving behaviors, distractions, and diverse environmental conditions. Factors such as driver inattention, reliance on smartphones, and varying traffic norms across regions significantly contribute to these incidents. As transportation systems become increasingly dynamic, there is an urgent need for adaptive, unobtrusive, and context-aware systems to mitigate risks and promote safer driving practices. Achieving true road safety requires a deep understanding and effective management of these challenges, highlighting the complex interactions between individual driving behavior, environmental conditions, and societal norms. In light of these concerns, this thesis investigates the issues arising from diverse driving micro-cultures. It explores how such variations can be leveraged to prevent unsafe driving practices and enhance road safety at a systemic level.

24/11/2024

New Paper Alert !!

Paper accepted at Expert Systems with Applications (ESWA). Heartiest congratulations to the authors.

Title: ProxyLabel: A Framework to Evaluate Techniques for Survey Fatigue Reduction Leveraging Auxiliary Modalities

Authors: Salma Mandi, Bivas Mitra

SUMMARY: Developing machine learning-based emotion inference models requires numerous ground truth labels, typically collected through self-reports, which can be burdensome and induce fatigue. Leveraging auxiliary modalities (e.g., facial expressions, speech signals) is a common approach to ease the burden that correlates with emotions; however, the absence of a common platform that offers the flexibility to explore different modalities and to select the key indicators for emotion prediction creates a major bottleneck. To address this, we develop a framework, ProxyLabel, which leverages auxiliary modalities to reduce the self-report collection burden for predicting emotion labels. In this framework, we implement multiple combinations of facial emotion recognition tools, voice clip analysis tools, and feature reduction approaches to demonstrate the versatility of the framework in facial expression and voice-based emotion self-report estimation. Evaluations involving 20 and 16 participants over 28-week and 22-week studies show ProxyLabel accurately estimates emotions from facial images (83% & 81% macro-F1 for valence and arousal) and voice clips (93% & 95% macro-F1), reducing self-report burden by 10% and 6%, respectively. Moreover, two context-sensitive applications are implemented to validate real-world efficacy of the framework.

Address

Department Of Computer Science And Engineering, IIT Kharagpur
Kharagpur
721302

Telephone

+919433124004

Website

http://cnerg.org/

Alerts

Be the first to know and let us send you an email when CNERG - Complex Networks Research Group, IIT Kharagpur posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Shortcuts

Want your university to be the top-listed University in Kharagpur?