Acceleration Consortium’s Post

Ph.D. students, here are some powerful strategies to help you write an unforgettable thesis in just 30 days. From utilizing voice-to-text technology to adopting existing theses as blueprints, these tips will guide you towards success. Develop a strong bond with your supervisor, simplify your methodology, and focus on minimalist, purposeful sentences. Remember, taking breaks and letting your mind wander can lead to breakthroughs. Let's conquer that thesis together! #PhD #ThesisWriting #AcademicSuccess

How can PhD students write an unforgettable PhD thesis in 30 days So, they’ve given you 1 month or you’re out on your ear? Remember… When you want to give up, help is closer than you think. When you want to give up, stop looking for miracles… Start looking for lessons. Change the conversation inside your head… stop running away Let’s go and get your thesis done… 1) Voice-to-text technology a] If you can speak, you can write—fast. Use voice-to-text tools, many are FREE. b] Record facts, ideas, and arguments–RAW… don’t edit, don’t interrupt the flow! c] Speak like you’re explaining your research to a friend over coffee. d] Got speakers block? Go walk your dog and tell this poor creature about your research problem, objective, etc. (Ch1). Your dog thinks you’re an idiot but who cares? e] Now edit and re-edit.   2) Adopt a Dr, don’t reinvent the wheel a] For over 23 years, I have told my registered students 1000s of times… Dive into YOUR university’s digital repository and dig up dozens of PhD theses in your field. b] Globally there are 1000s of them. They’re FREE blueprints! c] Find a thesis you can adopt and adapt to fit your topic. d] Then ask your AI friend to help you. e] Check out my LI post “How do PhD students surf the ChatGPT tsunami and not get caught?” https://lnkd.in/gSCSpSiX   3) Speed-read a] Speed-read Abstracts, Introductions, Conclusions, and especially Recommendations. Why? Because there are 100s of research gaps in Recommendations. b] Develop Chs 1-3 Intro, Literature Review, and Methodology using 3a] (above). c] Copy/paste in a specific colour, then after processing/paraphrasing change to black… avoids plagiarism.   4) Bond with your supervisor a] Some PhD students registered with me when they wanted to quit. Number 1 reason was supervisor issues. b] Surprisingly, most supervisors are human! c] Communicate often. Send weekly drafts, ask for specific feedback. Send personalized notes on birthdays, anniversaries, etc. Weekly updates = faster approvals. 5) Reverse engineer your methodology a] Kick off with the end in mind. What findings do you need to answer your research questions? b] Work backwards, keep it simple and direct. Avoid a complex methodological framework. c] Clearly align your methods with your research objectives.   6) Minimalist sentences a] Short, powerful sentences pack a punch. b] Be clear, be concise, avoid fluff. c] Every sentence has a purpose, so state, explain, rationale, argue, and/or summarize facts.   7) Distract your mind a] If you want to complete your thesis in 30 days, you’d better recharge yourself daily. b] Play hard, exercise, meditate, or let your mind mull over where you are in your thesis to kick-start your sub-conscious mind so it works on your thesis while you sleep. c] Let out the genius hiding inside—best ideas and solutions come when you’re not actively working on them. Want to register with me? rod_pallister@yahoo.co.uk

🎓 *Master's Thesis Journey Experience!* 🎓 🌟After months of dedication, research, and experimentation, I’m excited to share my Master’s thesis titled **TransNet: Fine-Tuned Vision Transformer Model for Offline Signature Verification** 🌟 This journey has been a rollercoaster of challenges and breakthroughs—countless late nights, refining algorithms, optimizing models, and pouring over data. With a focus on harnessing the power of deep learning, I developed a fine-tuned *TransNet*, a model that pushes the boundaries of accuracy in offline signature verification, showing significant improvements over existing state-of-arts methods. I would like to express my gratitude towards my guide Prof. Ajay Mittal from UIET Panjab University for his constant support, mentoring and exceptional guidance and motivation in the overall journey. Other than my guide, I am also grateful for receiving guidance and moral support from other faculty members of UIET Panjab University. Their research guidance and publications motivation, gradually enhanced my interest towards research. I’ve learned the importance of perseverance, collaboration, and continuous innovation. Following is my overall experience which I gained in the process:- ➡️ First and foremost important thing is to analyse the problem statement involved in the research. The problem statement often begins with a few words which we later stretch it as per our work requirement/needs. In my case it was initially simple being *Offline Signature Verification*, later it was formulated with thorough work analysis. ➡️ Secondly when you analyze your problem statement, it is very important to carry out Literature Survey/Review. It helps you to analyze the work that has been carried out in your research field. I did this survey portion with almost around 80-90 papers. ➡️ Third, you need to analyze your Research Gaps involved in your field of research. It means what type of improvements/novel work that you could produce in your field which might be considered as your own research work. ➡️ Fourth, here you need to carry out the Implementation part, which involves dataset(s) analysis, pre-prcoessing, features extraction, model's training/testing, classification etc. This phase is crucial and may take time depending upon how complex your task is. ➡️ Fifth, last stage which involve how you present your work in the form of publications as well as in a Theis format. This is yet an important and crucial step. The above enlightened experience is the generic one which might help you to understand how to effectively carry out research at master's level. If someone is interested in carrying out publications in their field of research, please feel free to reach out. Excited for the **next chapter** and eager to apply these learnings in the real world! 🚀 #MachineLearning #DeepLearning #VisionTransformer #OfflineSignature_Verification #MastersThesis

There is something unique and rare about sharing struggles. For one, many people are able to relate but more importantly, it can inspire others and equip them with the tools they need to find their own success. Olutoun S. Follow my page Pharmamed Spectrum Care #study #PhD #growth #lessons #struggles #support #student

Lessons from my MIT PhD struggles PhD is very hard. Everybody goes through tough times and comes out as a bruised (or mauled) warrior at the end. These are the biggest struggles I faced during my research. 1. Lab experiments: I came from a computational background at IITM and joined a purely experimental lab at MIT. On my day 1, I was overwhelmed by the equipment I saw in my lab. I could not fathom how I would ever become good at using all that equipments, let alone producing novel results. 2. Imposter syndrome: MIT had an orientation event for fresh grad students on the first day at a big auditorium. I saw students from all over the world, China, the USA, Europe, etc. Many of them were working on their laptops already, and I had not even met my advisor to discuss my project. I really felt like a nobody on that day. 3. Not having results: It was very hard to produce meaningful results in the first year. Not knowing the direction, I performed a lot of meaningless experiments. The struggle was so real to identify the right thing to do. 4. Shedding the “solution” mindset: Till the end of bachelors, we are taught to “solve” problems. Research is not just about finding the “correct” solutions. After all, what is a "correct" solution? It was very hard to get rid of this mindset after years of training to solve problems. 5. Uncertainty: The uncertainty of where my work and career were going was always there during a good part of my PhD. It was very hard to deal with. 6. Work-life balance: For a good half of my PhD, work-life balance didn’t exist. This came from a mindset that I had cultivated till then. Whenever I was relaxing, my mind took me on a guilt trip. Therefore, I was overworked and underproductive. 7. Deep work: This is very abstract, and I don’t know how to quantify it. But I felt during the first half of PhD, my thought process was shallow, and I could not get much deep work done. 8. Comparing with peers: This was a habit that helped me always- to compete in entrance exams and to secure good GPAs. But not in PhD. The journeys of 2 different researchers are on completely different trajectories. Comparing the no of papers, awards won, citations etc. is quite meaningless. 9. Coding: I became very good at experiments eventually. Felt very comfortable building any kind of setup. But that is when my comfort with coding took a hit. I was once very good at it. But then coding suddenly felt like a big task. I had to learn to get comfortable coding again during the last phase of PhD. 10. Communication: By far my greatest weakness during the early years of PhD. I used to speak fast and incoherently. Once, I was explaining an experiment I did to my advisor. My advisor could not understand what I was describing. Then my senior rephrased my words to make my advisor understand. It was amazing to witness. It was my work, and the senior student could explain it 10X better than I could. I will remember this one instance for the rest of my life.

Fully funded PhD studentship in Lifelong Robot Learning at the University of Southampton, UK The integration of Large Language Models (LLMs) into robot learning has recently demonstrated remarkable success. It became possible to ground LLMs in the physical context and solve long-horizon robotic tasks efficiently by querying LLMs to generate a sequence of natural language commands corresponding to pretrained skills that accomplish a given task. LLMs have also enabled few-shot adaptation to novel tools by generating task-agnostic tool descriptions for language-conditioned learning of manipulation skills. Long-horizon task learning and few-shot adaptation, supported by LLMs, are essential to lifelong robot learning. However, the current utilization of LLMs remains incompatible with the continual learning setting, wherein tasks are sequentially presented to the robot without a predefined task distribution. Moreover, the robot is expected to retain knowledge of previous tasks when learning new ones. This project aims to explore how LLMs can be harnessed to continuously acquire new skills to solve novel tasks as opposed to mastering a predefined and fixed set of tasks. In particular, methods for incrementally learning skill representations jointly from textual descriptions and spatio-temporal information of action sequences will be developed and evaluated on learning visuomotor robotic tasks in a household environment. You will join the School of Electronics and Computer Science which is ranked 1st in the UK for Electrical and Electronic Engineering (Guardian University Guide 2022) within the University of Southampton which is ranked in the top 1% of universities worldwide. The successful candidate must have a strong background in machine learning. Prior knowledge of reinforcement learning, robotics and training and finetuning LLMs is highly desirable. # Funding available for: - UK, EU and Horizon Europe Associated Countries - For exceptional overseas candidates, the fee difference between UK and overseas fees may be covered through other funding schemes, such as https://lnkd.in/es85V6vN # How to apply: Please send your pre-application package including CV, transcripts, sample publications (if any), two references and cover letter expressing your research vision and interest in the position to burhan.hafez@soton.ac.uk # Closing date: Applications will be considered in the order that they are received. The position will be considered filled when a suitable candidate has been identified. For more information, please visit: https://lnkd.in/ecKKyaAu

I’m really grateful to have experienced doctoral training. As I progress in my career, the soft skills I developed during my PhD, prior to me quitting, have been integral to the successes I’ve achieved in the industry roles I’ve held. Yes, having a PhD means you’ve earned an esteemed credential and, if you’re lucky, a few first-author papers to your name but, I’m so fortunate to have learned things like: 1. Critical thinking • Perhaps the most important skill I’ve carried with me beyond the academy: really framing a problem from hypothesis generation, through method selection and data interpretation to eventual completion. Or measured in industry KPIs, “to eventual tangible business value add.” 2. Reading primary literature • I used to consider this so onerous throughout undergrad and even at the start of grad school. I felt like I didn’t need to read papers. But dissecting methods and inferring conclusions is a skillset I’ve leaned on almost daily to stay abreast with Computational Biology & AI innovations. 3. Learning deeply, quickly • Being forced to collaborate with folks in otherwise unrelated fields meant I needed to become an expert in a particular research area unlike my own, rapidly. This has made me extremely adaptable over the years. 4. Functioning autonomously • No one holds your hand during a PhD. You’ve got to fumble around in the dark for a bit before you [hopefully] find your footing. You are responsible for spearheading your thesis work to completion. Though daunting, it’s made me confident in my own capabilities, even in uncharted territory. Dean Lee talks a lot about this stuff in pursuit of making Comp Bio careers accessible through his Figure One Lab initiative. He inspired this recollection of mine. Check him out. _________ #compbio #research #data #careers #phd

The progress in a Ph.D. program is indeed a highly non-linear function when plotted against time. This non-linearity arises due to the unique nature of research, personal development, and the various phases a Ph.D. student goes through. Here are some key phases and how they contribute to this non-linear progress: 1. Initial Phase (Exploration and Coursework): In the beginning, Ph.D. students typically spend significant time on coursework and exploring various research topics. This period involves: Literature Review: Reading and understanding the existing body of work. Coursework: Completing required classes to build a strong foundation. Identifying Research Gap: Finding a novel area or problem to focus on. During this phase, visible progress in terms of research output (papers published, experiments completed) is minimal. The curve is flat or shows slow progress. 2. Mid-Phase (Data Collection and Initial Experiments): Once a research direction is identified, the next phase involves: Designing Experiments: Planning and setting up experiments or studies. Data Collection: Gathering necessary data through experiments, surveys, or other methods. Initial Results: Analyzing early results, which may involve a lot of trial and error. This phase can be marked by bursts of progress interspersed with periods of stagnation as experiments need time to run, and initial results often require adjustments and further experimentation. 3. Development Phase (Analysis and Paper Writing); In this phase, students focus on: Data Analysis: Making sense of the collected data, which may involve complex statistical or computational methods. Writing Papers: Documenting findings in the form of research papers. Conferences and Feedback: Presenting at conferences and getting feedback for improvements. Progress becomes more visible as students produce tangible outputs (papers, presentations). However, the process is still non-linear due to the iterative nature of writing, revising, and resubmitting papers. 4. Final Phase (Dissertation Writing and Defense Preparation): The final phase involves: Dissertation Writing: Compiling the entire body of research into a dissertation. Defense Preparation: Preparing for the final defense, which includes summarizing and defending the research work. This phase is characterized by intense writing and revision efforts, leading up to a final, significant spike in progress as the dissertation is submitted and defended. When combining all these phases, the overall progress in a Ph.D. looks like a highly non-linear function with periods of slow progress, sudden bursts of advancement, and occasional plateaus or even setbacks. The journey is marked by a mixture of learning, experimentation, and output creation, each contributing to the non-linear nature of the progress curve. #ResearchJourney #PhD #PhDScholar #PhDProgress #PhDLife #LearningCurve

🚨 𝗧𝗵𝗲 𝗦𝗣𝗘𝗖𝗜𝗘𝗦 𝗦𝗰𝗵𝗼𝗹𝗮𝗿𝘀𝗵𝗶𝗽 𝗰𝗮𝗹𝗹 𝗳𝗼𝗿 𝗰𝗮𝗻𝗱𝗶𝗱𝗮𝘁𝗲𝘀 𝗶𝘀 𝘀𝘁𝗶𝗹𝗹 𝗼𝗽𝗲𝗻! This year, up to three scholarships will be awarded, offering the opportunity to spend three months at a host institution. See more here to learn how to apply: https://lnkd.in/d-JU5F3v 🚨 We've been awarding scholarships since 2020. Are you curious about the amazing work that has resulted from these opportunities? Here are the recipients from 2020. 🚁 Muhammad Ali Nayeem, PhD • Host: Ahmed Kheiri, University of Lancaster, UK • Publication: GECCO’21 - https://lnkd.in/dWBDTFiY • Project: North West Air Ambulance keeps NW England safe with 3 teams & 3 helicopters from 2 bases! This work models their operations using a matheuristic approach to see how base locations and team setups affect service quality. 🚂 Bianca Pascariu • Host: Paola Pellegrini, University Gustave Eiffel, France • Publication: IFAC’2021 - https://lnkd.in/d7sYybB2 • Project: Tackling the real-time Railway Traffic Management Problem (rtRTMP), Bianca's work focuses on minimizing disturbances in railway networks. By using Ant Colony Optimization (ACO) and integer linear programming, her research demonstrates superior performance in solving large instances of the Train Routing Selection Problem (TRSP). 🔍 Adam Katona • Hosts: Nuno Lourenço and Penousal Machado, University of Coimbra, Portugal • Publication: Evostar 2021 - https://lnkd.in/dub5-YVA • Project: Adam's research explores indirect encoding in machine learning and evolutionary computation. His work demonstrates that while indirect encoding can sometimes underperform in normal learning, optimizing for adaptability using the MAML algorithm can allow indirect encoding to outperform direct encoding, particularly in few-shot image classification tasks. 📊 Yuri Lavinas • Host: Gabriela Ochoa, University of Stirling, UK • Publication: Evostar 2022 - https://lnkd.in/dt-RsPD3 • Yuri extended a network-based tool called search trajectory networks (STNs) to model the behavior of multiobjective evolutionary algorithms (MOEAs). His findings highlight how STNs can improve our understanding of the search dynamics of MOEA/D and NSGA-II algorithms across various benchmark problems. 🖋️ Sérgio M. Rebelo • Host: JJ Merelo, University of Granada, Spain • Publication: Evomusart 2024 - https://lnkd.in/dr5paTCJ • Project: Sérgio proposed heuristic metrics for evaluating typographic design, focusing on legibility, aesthetics, and semantic features. His research integrates a constrained evolutionary approach and emotion recognition to generate and analyze typographic posters, aiming to enhance design evaluation methods.

Top 10 Carolina Science Online Secrets to Unlocking the Key Features: Beginners are welcomed into the world of Carolina Science Online, where a wealth of information is just waiting to be discovered. This website provides a dynamic learning experience beyond the conventional bounds of science education, with engaging articles on the latest scientific breakthroughs and immersive virtual activities. At first, sifting through the enormous amount of resources may seem overwhelming, but don’t worry-Carolina Science Online has an easy-to-use interface with helpful search tools and categories to make navigating its depths a breeze. This online resource opens doors to a world of scientific exploration that is boundless for anyone to start their scientific path or satisfy their curiosity. Read More... https://lnkd.in/d2yDdiEU