skip to content

CambPlants

A networking organisation for plants-related research and impact
 
Subscribe to http://talks.cam.ac.uk/show/rss/49845 feed
A list of talks of interest to the Cambridge Centre for Data-Driven Discovery (C2D3) network, pulled together from around the University.
Updated: 1 hour 20 min ago

Wed 31 Aug 16:30: Statistics Clinic Summer 2022 III The clinic takes place in MR5 at the Centre for Mathematical Sciences.

Mon, 01/08/2022 - 12:12
Statistics Clinic Summer 2022 III

If you would like to participate, please fill in the following form. The deadline for signing up for a session is 12pm on Monday the 29th of August. Subject to availability of members of the Statistics Clinic team, we will confirm your in-person or remote appointment.

This event is open only to members of the University of Cambridge (and affiliated institutes). Please be aware that we are unable to offer consultations outside clinic hours.

The clinic takes place in MR5 at the Centre for Mathematical Sciences.

Add to your calendar or Include in your list

Mon 22 Aug 15:30: BSU Seminar: "Genome-wide genetic models for association, heritability analyses and prediction" This will be a free hybrid seminar. To register to attend virtually, please click here: https://us02web.zoom.us/meeting/register...

Wed, 27/07/2022 - 09:50
BSU Seminar: "Genome-wide genetic models for association, heritability analyses and prediction"

Although simultaneous analysis of genome-wide SNPs has been popular for over a decade, the problems posed by more SNPs than study participants (more parameters than data points), and correlations among the SNPs, have not been adequately overcome so that almost all published genome-wide analyses are suboptimal. While there has been much attention paid to the shape of prior distributions for SNP effect sizes, we argue that this attention is misplaced. We focus on what we call the “heritability model”: a low-dimensional model for the expected heritability at each SNP , which is key to both individual-data and summary-statistic analyses. The 1-df uniform heritability model has been implicitly adopted in a wide range of analyses. Replacing it with better heritability models, using predictors based on allele frequency, linkage disequilibrium and functional annotations, leads to substantial improvements in estimates of heritability and selection parameters over traits, and over genome regions, as well as improvements in gene-based association testing and prediction. Key collaborators Doug Speed, Aarhus, Denmark and Melbourne PhD student Anubhav Kaphle.

This will be a free hybrid seminar. To register to attend virtually, please click here: https://us02web.zoom.us/meeting/register/tZcpcuipqTMuGtP_pIJrVQDLQGbP3ndqAh-N

Add to your calendar or Include in your list

Sun 20 Nov 17:00: A Distance Function based Cascaded Neural Network for accurate Polyps Segmentation and Classification

Wed, 20/07/2022 - 08:20
A Distance Function based Cascaded Neural Network for accurate Polyps Segmentation and Classification

In clinical practice, it is often difficult to locate and measure the size of polyps accurately for the follow-up surgical operation decision. In this paper, based on the position constraint between the primary organ and polyps boundary, we propose a U-Net based cascaded neural network for the joint segmentation of the organ of interest and polyps. The constraint on their position relation is further imposed by adding a narrow-band distance function and complimentary dice function to the loss function. Through a series of comparisons and ablation study, the proposed method with the cascaded network architecture and the additional loss functions was validated on an in-house dataset for gallbladder polyps segmentation and classification. It has been demonstrated that the proposed method achieved a significant improvement over conventional U-Net, U-Net++ etc.. Eventually, the pathological type classification based on the segmented polys shows 30% higher accuracy compared to those conventional ResNet based result

Add to your calendar or Include in your list

Tue 26 Jul 14:00: Analysis by synthesis for interpretable image collection analysis

Tue, 19/07/2022 - 15:30
Analysis by synthesis for interpretable image collection analysis

Abstract

I will present our recent work on analyzing the content of image collections by learning a simple prototype-based model of images. I will start by introducing the idea and framework of Deep Transformation Invariant image analysis in the case of image clustering [1], where that a simple modification of the standard K-means algorithm can lead to state of the art image clustering, while computing distances in pixel space and being easy to interpret. I will then show how the idea can be extended to perform object recovery [3], decomposing every image in a collection into layers derived from a small set of image prototypes. This can be applied to real world data, such as collection of Instagram images, and provide models and segmentation of repeated objects. Finally, I will explain how a similar idea can be used to perform single view reconstruction from a categorical image collection without any supervision.

[1] Deep Transformation-Invariant Clustering, T. Monnier, T. Groueix, M. Aubry, NeurIPS 2020, link

[2] Unsupervised Layered Image Decomposition into Object Prototypes, T. Monnier, E. Vincent, J. Ponce, M. Aubry, ICCV 2021 , link

[3] Share With Thy Neighbors: Single-View Reconstruction by Cross-Instance Consistency, T. Monnier, M. Fisher, A. Efros, M. Aubry, ECCV 2022 , link

Bio

Mathieu Aubry is a tenured researcher in the Imagine team of Ecole des Ponts ParisTech. His work is mainly focussed on Computer Vision and Deep Learning, and their intersection with Computer Graphics, Machine Learning, and Digital Humanities. His PhD on 3D shapes representations obtained in 2015 at ENS was co-advised by Josef Sivic (INRIA) and Daniel Cremers (TUM). In 2015, he spent a year working as a postdoc with Alexei Efros in UC Berkeley.

Location

This seminar will be run as a hybrid event. Attendees can join in person by coming to the MIL Meeting Room in Baker Building (CB2 1PZ). You can also join via Zoom, using the provided link.

Google Calendar for Future Seminars: https://calendar.google.com/calendar/u/0?cid=c2pjcHN0YXM2N3QyMWU3c2FqNjBqNWNiYXNAZ3JvdXAuY2FsZW5kYXIuZ29vZ2xlLmNvbQ

Add to your calendar or Include in your list

Mon 22 Aug 15:30: BSU Seminar: "Genome-wide genetic models for association, heritability analyses and prediction" This will be a free hybrid seminar. To register to attend virtually, please click here: https://us02web.zoom.us/meeting...

Tue, 19/07/2022 - 13:51
BSU Seminar: "Genome-wide genetic models for association, heritability analyses and prediction"

Although simultaneous analysis of genome-wide SNPs has been popular for over a decade, the problems posed by more SNPs than study participants (more parameters than data points), and correlations among the SNPs, have not been adequately overcome so that almost all published genome-wide analyses are suboptimal. While there has been much attention paid to the shape of prior distributions for SNP effect sizes, we argue that this attention is misplaced. We focus on what we call the “heritability model”: a low-dimensional model for the expected heritability at each SNP , which is key to both individual-data and summary-statistic analyses. The 1-df uniform heritability model has been implicitly adopted in a wide range of analyses. Replacing it with better heritability models, using predictors based on allele frequency, linkage disequilibrium and functional annotations, leads to substantial improvements in estimates of heritability and selection parameters over traits, and over genome regions, as well as improvements in gene-based association testing and prediction. Key collaborators Doug Speed, Aarhus, Denmark and Melbourne PhD student Anubhav Kaphle.

This will be a free hybrid seminar. To register to attend virtually, please click here: https://us02web.zoom.us/meeting/register/tZAsdeyopj0uGdYCZRIOFnl1wH5y0cU3amO-

Add to your calendar or Include in your list

Fri 28 Oct 17:00: Approximate Equivariance SO(3) Needlet Convolution

Mon, 18/07/2022 - 18:12
Approximate Equivariance SO(3) Needlet Convolution

This work focuses on the development of a rotation-invariant needlet convolution for rotation group SO(3) to distill multiscale information of spherical signals. The spherical needlet transform is generalized from $\sS^2$ onto the SO(3) group, which decomposes a spherical signal to approximate and detailed spectral coefficients by a set of tight framelet operators. The spherical signal during the decomposition and reconstruction achieves rotation invariance. Based on needlet transforms, we form a Needlet approximate Equivariance Spherical CNN (NES) with multiple SO(3) needlet convolutional layers. The network establishes a powerful tool to extract geometric-invariant features of spherical signals. The model allows sufficient network scalability with multi-resolution representation. A robust signal embedding is learned with wavelet shrinkage activation function, which filters out redundant high-pass representation while maintaining approximate rotation invariance. The NES achieves state-of-the-art performance for quantum chemistry regression and Cosmic Microwave Background (CMB) delensing reconstruction, which shows great potential for solving scientific challenges with high-resolution and multi-scale spherical signal representation.

Add to your calendar or Include in your list

Tue 27 Sep 13:00: ACMP: Allen-Cahn Message Passing for Graph Neural Networks with Particle Phase Transition

Mon, 18/07/2022 - 15:26
ACMP: Allen-Cahn Message Passing for Graph Neural Networks with Particle Phase Transition

Neural message passing is a basic feature extraction unit for graph-structured data that takes account of the impact of neighboring node features in network propagation from one layer to the next. We model such process by an interacting particle system with attractive and repulsive forces and the Allen-Cahn force arising in the modeling of phase transition. The system is a reaction-diffusion process which can separate particles to different clusters. This induces an Allen-Cahn message passing (ACMP) for graph neural networks where the numerical iteration for the solution constitutes the message passing propagation. The mechanism behind ACMP is phase transition of particles which enables the formation of multi- clusters and thus GNNs prediction for node classification. ACMP can propel the network depth to hundreds of layers with theoretically proven strictly positive lower bound of the Dirichlet energy. It thus provides a deep model of GNNs which circumvents the common GNN problem of oversmoothing. Experiments for various real node classification datasets, with possible high homophily difficulty, show the GNNs with ACMP can achieve state of the art performance with no decay of Dirichlet energy.

Joint work with Yuelin Wang (SJTU), Kai Yi (UNSW), Xinliang Liu (KAUST) and Shi Jin (SJTU).

Add to your calendar or Include in your list

Wed 13 Jul 16:30: Statistics Clinic Summer 2022 I The clinic takes place in MR2 at the Centre for Mathematical Sciences.

Wed, 13/07/2022 - 16:16
Statistics Clinic Summer 2022 I

If you would like to participate, please fill in the following form. The deadline for signing up for a session is 12pm on Monday the 11th of July. Subject to availability of members of the Statistics Clinic team, we will confirm your in-person or remote appointment.

This event is open only to members of the University of Cambridge (and affiliated institutes). Please be aware that we are unable to offer consultations outside clinic hours.

The clinic takes place in MR2 at the Centre for Mathematical Sciences.

Add to your calendar or Include in your list

Thu 14 Jul 14:00: Image Quality Metrics at the Time of Deep Learning The talk can also be attended via zoom. Join Zoom Meeting https://cl-cam-ac-uk.zoom.us/j/92211133320?pwd=dHJqR3lnYzNCN0ZxYnIzRUNZUnRMdz09 ID: 92211133320 passcode: 054090

Wed, 13/07/2022 - 16:10
Image Quality Metrics at the Time of Deep Learning

In this talk, we present the current status of image quality metrics in this deep learning age. From this, we will analyze a possible solution to speed-up classic perceptual image quality metrics, and how to move into the no-reference domain. We will show applications why no reference metrics may be useful for image quality assessment.

The talk can also be attended via zoom. Join Zoom Meeting https://cl-cam-ac-uk.zoom.us/j/92211133320?pwd=dHJqR3lnYzNCN0ZxYnIzRUNZUnRMdz09 ID: 92211133320 passcode: 054090

  • Speaker: Francesco Banterle, Visual Computing Lab, ISTI-CNR, Italy
  • Thursday 14 July 2022, 14:00-15:00
  • Venue: SS03.
  • Series: Rainbow Group Seminars; organiser: am2806.

Add to your calendar or Include in your list

Wed 13 Jul 16:30: Statistics Clinic Summer 2022 I The clinic takes place in MR11 at the Centre for Mathematical Sciences.

Wed, 13/07/2022 - 11:06
Statistics Clinic Summer 2022 I

If you would like to participate, please fill in the following form. The deadline for signing up for a session is 12pm on Monday the 11th of July. Subject to availability of members of the Statistics Clinic team, we will confirm your in-person or remote appointment.

This event is open only to members of the University of Cambridge (and affiliated institutes). Please be aware that we are unable to offer consultations outside clinic hours.

The clinic takes place in MR11 at the Centre for Mathematical Sciences.

Add to your calendar or Include in your list

Wed 03 Aug 16:30: Statistics Clinic Summer 2022 II The clinic takes place in MR5 at the Centre for Mathematical Sciences.

Mon, 11/07/2022 - 10:37
Statistics Clinic Summer 2022 II

If you would like to participate, please fill in the following form. The deadline for signing up for a session is 12pm on Monday the 1st of August. Subject to availability of members of the Statistics Clinic team, we will confirm your in-person or remote appointment.

This event is open only to members of the University of Cambridge (and affiliated institutes). Please be aware that we are unable to offer consultations outside clinic hours.

The clinic takes place in MR5 at the Centre for Mathematical Sciences.

Add to your calendar or Include in your list

Thu 14 Jul 14:00: BSU Seminar: "Challenges in risk prediction using routinely collected health data" This will be a free online seminar. To register, click here: https://us02web.zoom.us/meeting/register/tZEkdOupqDspHtK1D30gXlBykdbIvLV8DMdH

Mon, 11/07/2022 - 09:39
BSU Seminar: "Challenges in risk prediction using routinely collected health data"

Identifying who is at highest risk of severe consequences of COVID -19, such as hospitalisation or death, is an important component of any policy response. One of the best sources of information on who experiences these events is linked electronic health record data. Using these data to identify high-risk patients, however, raises a number of methodological challenges. This talk discusses the use of linked primary care data within the OpenSAFELY platform to predict risk of COVID -19 mortality and the challenges of doing so.

This will be a free online seminar. To register, click here: https://us02web.zoom.us/meeting/register/tZEkdOupqDspHtK1D30gXlBykdbIvLV8DMdH

Add to your calendar or Include in your list

Thu 21 Jul 15:00: cISP: A Speed-of-Light Internet Service Provider

Mon, 11/07/2022 - 09:20
cISP: A Speed-of-Light Internet Service Provider

Low latency is a requirement for a variety of interactive network applications. The Internet, however, is not optimized for latency. We thus explore the design of wide-area networks that move data at nearly the speed of light in vacuum. Our cISP design augments the Internet’s fiber with free-space microwave wireless connectivity over paths very close to great-circle paths. cISP addresses the fundamental challenge of simultaneously providing ultra-low latency while accounting for numerous practical factors ranging from transmission tower availability to packet queuing, achieving mean latencies within 5% of that achievable using great-circle paths at the speed of light. Further, using experiments conducted on a nearly-speed-of-light algorithmic trading network, together with an analysis of trading data at its end points, we show that microwave networks are reliably faster than fiber networks even in inclement weather. We provide estimates showing the economic value of such networks would substantially exceed their expense. Finally, we will discuss how this work can influence future directions using a combination of multiple communication channels—low latency and high bandwidth—such as with 5g wireless.

Bio: Brighten Godfrey is a professor in the Department of Computer Science at the University of Illinois at Urbana-Champaign, and a technical director at VMware. He co-founded and served as CTO of network verification pioneer Veriflow through its 2019 acquisition by VMware. He received his Ph.D. at UC Berkeley in 2009. His research interests lie in the design of networked systems and algorithms. He is a winner of the ACM SIGCOMM Rising Star Award, the Sloan Research Fellowship, and the National Science Foundation CAREER Award, and has chaired several conferences including ACM HotNets 2014, the Symposium on SDN Research 2016, and ACM SIGCOMM 2022 .

Add to your calendar or Include in your list

Thu 18 Aug 15:00: Achieving Consistent Low Latency for Wireless Real-Time Communications with the Shortest Control Loop

Sun, 10/07/2022 - 08:21
Achieving Consistent Low Latency for Wireless Real-Time Communications with the Shortest Control Loop

Real-time communication (RTC) applications like video conferencing or cloud gaming require consistent low latency to provide a seamless interactive experience. However, wireless networks including WiFi and cellular, albeit providing a satisfactory median latency, drastically degrade at the tail due to frequent and substantial wireless bandwidth fluctuations. We observe that the control loop for the sending rate of RTC applications is inflated when congestion happens at the wireless access point (AP), resulting in untimely rate adaption to wireless dynamics. Existing solutions, however, suffer from the inflated control loop and fail to quickly adapt to bandwidth fluctuations. In this paper, we propose Zhuge, a pure wireless AP based solution that reduces the control loop of RTC applications by separating congestion feedback from congested queues. We design a Fortune Teller to precisely estimate per-packet wireless latency upon its arrival at the wireless AP. To make Zhuge deployable at scale, we also design a Feedback Updater that translates the estimated latency to comprehensible feedback messages for various protocols and immediately delivers them back to senders for rate adaption. Trace-driven and real-world evaluation shows that Zhuge reduces the ratio of large tail latency and RTC performance degradation by 17% to 95%.

Speaker Bio: Zili is a 3rd-year PhD student in Tsinghua University. His current research interest focuses on real-time video communications. He has published several papers in SIGCOMM / NSDI and received the Microsoft Research Asia PhD Fellowship, Gold Medal of SIGCOMM 2018 Student Research Competition, and two best paper awards.

Add to your calendar or Include in your list

Thu 18 Aug 15:00: Achieving Consistent Low Latency for Wireless Real-Time Communications with the Shortest Control Loop

Sun, 10/07/2022 - 08:20
Achieving Consistent Low Latency for Wireless Real-Time Communications with the Shortest Control Loop

Real-time communication (RTC) applications like video conferencing or cloud gaming require consistent low latency to provide a seamless interactive experience. However, wireless networks including WiFi and cellular, albeit providing a satisfactory median latency, drastically degrade at the tail due to frequent and substantial wireless bandwidth fluctuations. We observe that the control loop for the sending rate of RTC applications is inflated when congestion happens at the wireless access point (AP), resulting in untimely rate adaption to wireless dynamics. Existing solutions, however, suffer from the inflated control loop and fail to quickly adapt to bandwidth fluctuations. In this paper, we propose Zhuge, a pure wireless AP based solution that reduces the control loop of RTC applications by separating congestion feedback from congested queues. We design a Fortune Teller to precisely estimate per-packet wireless latency upon its arrival at the wireless AP. To make Zhuge deployable at scale, we also design a Feedback Updater that translates the estimated latency to comprehensible feedback messages for various protocols and immediately delivers them back to senders for rate adaption. Trace-driven and real-world evaluation shows that Zhuge reduces the ratio of large tail latency and RTC performance degradation by 17% to 95%.

Speaker Bio: Zili is a 3rd-year PhD student in Tsinghua University. His current research interest focuses on real-time video communications. He has published several papers in SIGCOMM / NSDI and received the Microsoft Research Asia PhD Fellowship, Gold Medal of SIGCOMM 2018 Student Research Competition, and two best paper awards.

Add to your calendar or Include in your list

Wed 13 Jul 12:00: An Introduction to Data and Commercialisation

Fri, 08/07/2022 - 12:55
An Introduction to Data and Commercialisation

Hosted by Cambridge Enterprise and Cambridge Centre for Data-Driven Discovery (C2D3).

There are two parts to this webinar:

Part 1: ‘Future Proofing your Data Project’

Building a foundation for dissemination and impact into your research project can feel overwhelming when you’re dealing in data. We look at some practical steps you can take to manage data when you have commercialisation in mind, and talk about when and how to find support from across the University.

Managing data with future commercialisation in mind Basics of ‘protecting data’ and IP and an introduction to commercial licensing Using third party and open source materials Do’s and don’ts with health and personal data *When things go wrong!

Part 2: ‘Creating Impact through Data Commercialisation’

We’ll introduce some case studies of data commercialisation from Cambridge and beyond. Data commercialisation uses diverse models including consultancy, licensing and company creation. How can you devise a pathway to maximise your impact and create financial sustainability for your project beyond grant funding?

Speakers will include:

Dr Terry Parlett, Dr Emma Salgård Cunha and Dr Sian Fogden

Webinar details:

Date: Wednesday 13th July 2022 Time: 12:00 – 13:30 Platform: Microsoft Teams

Add to your calendar or Include in your list

Tue 12 Jul 14:00: Claim-Dissector: An Interpretable Fact-Checking System with Joint Re-ranking and Veracity Prediction

Wed, 06/07/2022 - 13:50
Claim-Dissector: An Interpretable Fact-Checking System with Joint Re-ranking and Veracity Prediction

Abstract:

We present Claim-Dissector: a novel latent variable model for fact-checking and fact-analysis, which given a claim and a set of retrieved provenances allows learning jointly (i) what are the provenances relevant to this claim (ii) what is the veracity of this claim. We show that our system achieves state-of-the-art results on FEVER comparable to two-stage systems often used in traditional fact-checking pipelines, while using significantly less parameters and computation. Our analysis shows that proposed approach further allows to learn not just which provenances are relevant, but also which provenances lead to supporting and which toward denying the claim, without direct supervision. This not only adds interpretability, but also allows to detect claims with conflicting evidence automatically. Furthermore, we study whether our model can learn fine-grained relevance cues while using coarse-grained supervision. We show that our model can achieve competitive sentence-recall while using only paragraph-level relevance supervision. Finally, traversing towards the finest granularity of relevance, we show that our framework is capable of achieving strong token-level interpretability. To do this, we present a new benchmark focusing on token-level interpretability ― humans annotate tokens in relevant provenances they considered essential when making their judgement. Then we measure how similar are these annotations to tokens our model is focusing on. Our code, dataset and demo will be released online.

Bio:

Martin Fajčík (read as Fay-Cheek) is a PhD candidate in Natural Language Processing from Knowledge Technology Research Group active at FIT -BUT in Brno, Czech Republic, advised by prof. Pavel Smrž (ž is read like j in french “Jean”). From 2021, he also works as a research assistant in IDIAP research institute based in Martigny, Switzerland. His PhD work is focusing on open-domain knowledge processing, mainly in question answering and fact-checking. He enjoys a good hikes and an informal discussions over tea.

Add to your calendar or Include in your list

Fri 08 Jul 11:00: Synthetics with Digital Humans

Mon, 04/07/2022 - 21:36
Synthetics with Digital Humans

Abstract

Nowadays, collecting the right dataset for machine learning is often more challenging than choosing the model. We address this with photorealistic synthetic training data – labelled images of humans made using computer graphics. With synthetics we can generate clean labels without annotation noise or error, produce labels otherwise impossible to annotate by hand, and easily control variation and diversity in our datasets. I will show you how synthetics underpins our work on understanding humans, including how it enables fast and accurate 3D face reconstruction, in the wild.

Bio

Dr. Erroll Wood is a Staff Software Engineer at Google, working on Digital Humans. Previously, he was a member of Microsoft’s Mixed Reality AI Lab, where he worked on hand tracking for HoloLens 2, avatars for Microsoft Mesh, synthetic data for face tracking, and Holoportation. He did his PhD at the University of Cambridge, working on gaze estimation.

Google Calendar for Future Seminars: https://calendar.google.com/calendar/u/0?cid=c2pjcHN0YXM2N3QyMWU3c2FqNjBqNWNiYXNAZ3JvdXAuY2FsZW5kYXIuZ29vZ2xlLmNvbQ

Add to your calendar or Include in your list