Search Marsden awards 2008–2017
Search awarded Marsden Fund grants 2008–2017
Fund Type: Marsden Fund
Category: Standard
Year Awarded: 2014
Title: Auckland lakes as climate dipsticks: producing unique insights into the nature and drivers of the past 117,000 years of climate change
Recipient(s): Associate Professor PC Augustinus | PI | The University of Auckland
Professor F Sirocko | PI | Johannes Gutenberg University
Dr M Danišík | AI | The University of Waikato
Professor IF Snowball | AI | Uppsala University
Dr TM Shanahan | AI | The University of Texas at Austin
Associate Professor PA Shane | AI | The University of Auckland
Professor AK Schmitt | AI | University of California, Los Angeles
Professor RM Newnham | AI | Victoria University of Wellington
Dr HM Roberts | AI | Aberystwyth University
Public Summary: The response of past terrestrial ecosystems to rapid climate change is central to the debate surrounding the consequences of future climate change in a warming world. However, although much has been learned regarding the mechanisms of abrupt climate change, it is unclear how these events are generated and transmitted. Part of the problem lies in the paucity of accurately and precisely dated high-resolution records of climate change needed to identify the degree of synchrony between these events throughout the global climate system. The episodes of rapid change over the past 117,000 years form part of the solution, although they best expressed in climate records from the polar and tropical regions. The Southern Hemisphere mid-latitudes are a data sparse and poorly-understood part of the global climate system, although our Auckland volcanic lake sediments will enable us to produce a unique, continuous and high-resolution record of past climate variability spanning the period of interest. We will use a novel approach involving a range of quantitative sedimentological, geochemical and biotic estimates of past climate with the development of robust age models. The output will be an unparalleled record of the nature, timing, rate and consequences of past climate change from the Southwest Pacific.
Total Awarded: $800,000
Duration: 3
Host: The University of Auckland
Contact Person: Associate Professor PC Augustinus
Panel: ESA
Project ID: 14-UOA-040
Fund Type: Marsden Fund
Category: Fast-Start
Year Awarded: 2016
Title: Automatic Design of Heuristics for Dynamic Arc Routing Problem with Genetic Programming
Recipient(s): Dr Y Mei | PI | Victoria University of Wellington
Professor J Branke | AI | University of Warwick
Public Summary: The Arc Routing Problem (ARP) is important in fields such as logistics and transportation. It aims to find minimum-cost vehicle routes to serve arcs of a graph. In real life, unpredicted events (e.g. traffic jam/accidents) can occur while executing the routes, leading to dynamic ARP. Traditional optimisation approaches cannot effectively respond to real-time changes in dynamic ARP.
Routing policy is a promising heuristic approach to dynamic ARP that can immediately determine the next destination for a vehicle once it finishes the current service. The effectiveness of a routing policy depends on a variety of interdependent factors. Therefore, manually designing effective routing policies is difficult. Using ideas from biological evolution, genetic programming (GP) has achieved success in automatically designing heuristics for many dynamic problems.
This project aims to propose novel GP techniques to automatically design routing policies in dynamic ARP. We will develop new strategies to address issues caused by multiple kinds of interacting decisions (assigning arcs to vehicles and picking the next arc), huge search space and multiple conflicting objectives (e.g. effectiveness and interpretability). We expect the proposed algorithms to provide effective and understandable routing policies for a wide range of dynamic ARP problems by discovering and reusing common knowledge.
Total Awarded: $300,000
Duration: 3
Host: Victoria University of Wellington
Contact Person: Dr Y Mei
Panel: MIS
Project ID: 16-VUW-079
Fund Type: Marsden Fund
Category: Fast-Start
Year Awarded: 2014
Title: Automatically deriving the model of digital ink sketches
Recipient(s): Dr RV Blagojevic | PI | Massey University
Associate Professor BE Plimmer | AI | The University of Auckland
Public Summary: Diagrams are critical to many areas including design, technical documentation, and education, as they have the ability to represent ideas or information in a visual and unconstrained manner. While computer supported diagramming is superficially attractive, to be useful it needs significant improvement. Enabling computer understanding of hand-drawn diagrams (such as via digital pens) is one way to achieve this. With recognition algorithms to automatically understand sketches, advanced functions can be added to software. The key is accurate recognition of the sketch. The overall aim of this research is to automatically infer the meaning — a conceptual model — of digital ink sketches. Understanding the high-level meaning of a sketch is an important part of sketch recognition; this currently requires experts to define the sketch semantics, which is incredibly tedious. In addition, current approaches to grouping sketched components, an important precursor to high-level understanding, lack effectiveness and scalability. This project aims to automatically and robustly perform high-level recognition and grouping. These advances in sketch recognition have the potential for wider implications in related areas such as annotation, touch, gesture, and scene recognition.
Total Awarded: $300,000
Duration: 3
Host: Massey University
Contact Person: Dr RV Blagojevic
Panel: MIS
Project ID: 14-MAU-025
Fund Type: Marsden Fund
Category: Standard
Year Awarded: 2017
Title: AviaNZ: Making Sure New Zealand Birds Are Heard
Recipient(s): Professor SR Marsland | PI | Massey University Manawatu
Associate Professor IC Castro | AI | Massey University Manawatu
Public Summary: New Zealand has amazing birdlife: nocturnal parrots, birds that can't fly, and birds that turn up after 50 years of being thought extinct. Unfortunately, many native species require wildlife management programmes, and they are hard to monitor: they are often well-camouflaged or nocturnal. How can you discern what works in wildlife management if you don't know how many animals there are and how the number changes?
Most birds make sound, and so there is a real need for a system that can detect, recognise, analyse, and infer bird populations using automated sound recorders. However, data from these recorders is noisy and extremely variable in factors such as volume. In this project we will work to reliably detect and recognise birdsong from these recordings, and estimate the abundance of birds from the number of calls, by developing mathematical and computational tools to analyse sound, and combining them with ecological experiments to understand how calls relate to population estimates. This inter-disciplinary project will combine fundamental and practical work in theoretical and experimental science to develop sound solutions in both areas. Our work will be made publicly available through our AviaNZ software platform, which is already used by conservation groups across the country.
Total Awarded: $880,000
Duration: 3
Host: Massey University Manawatu
Contact Person: Professor SR Marsland
Panel: EIS
Project ID: 17-MAU-154
Fund Type: Marsden Fund
Category: Standard
Year Awarded: 2014
Title: Axioms and algorithms for multi-winner elections
Recipient(s): Professor AM Slinko | PI | The University of Auckland
Dr MC Wilson | AI | The University of Auckland
Dr Pritchard | AI | The University of Auckland
Dr P Faliszewski | AI | AGH Institute of Science and Technology
Public Summary: Multi-winner elections are even more ubiquitous than single-winner ones, but much less studied. One reason is that there is a variety of models coexisting within this concept. We aim to classify the important types of multi-winner elections and, for each type, formulate the desirable normative properties and investigate any incompatibilities between these axioms. It is a commonplace for single-winner elections that not all desirable properties are compatible; this became clear from the famous Arrow (1951) and Gibbard-Satterthwaite (1973-75) impossibility theorems and further (and stronger) impossibility theorem of Slinko and White (2008). So trade-offs are inevitable and this makes the work of electoral designers complicated. However in the multi-winner case we do not even know what the trade-offs are. The emerging trade-offs will be studied theoretically and experimentally using probabilistic modelling methods developed by us in recent simulations of the NZ electoral system. Some methods of fully proportional representation e.g., those suggested by Chamberlin-Courant (1983) and Monroe (1995) have many attractive features but are proven to be NP-complete to implement. We will look for approximation algorithms for these two rules which also have nice normative properties. Our research will contribute to the design of new and effective voting systems.
Total Awarded: $535,000
Duration: 3
Host: The University of Auckland
Contact Person: Professor AM Slinko
Panel: MIS
Project ID: 14-UOA-254
Fund Type: Marsden Fund
Category: Standard
Year Awarded: 2008
Title: Back to the future: how the brain uses the past to imagine the future
Recipient(s): Dr DR Addis | PI | Harvard University
Prof MC Corballis | AI | The University of Auckland
Prof DL Schacter | AI | Harvard University
Public Summary: Research on memory has traditionally focused on past events, but recent work, including our own, examines how memory for past events enables us to imagine and plan the future. This work has revealed a core neural network that supports both remembering the past and simulating the future, but is more intensively engaged in future than past events. We plan to use neuroimaging (fMRI) to identify the mental processes that drive this increased activity during the simulation of future events. It will provide a more comprehensive understanding of how the brain draws upon details in our memories to simulate our futures.
Total Awarded: $590,222
Duration: 3
Host: The University of Auckland
Contact Person: Dr DR Addis
Panel: EHB
Project ID: 08-UOA-176
Fund Type: Marsden Fund
Category: Standard
Year Awarded: 2010
Title: Bacterial breakdown in a high-CO2 ocean
Recipient(s): Dr CS Law | PI | NIWA - The National Institute of Water and Atmospheric Research Ltd
Dr E Maas | PI | NIWA - The National Institute of Water and Atmospheric Research Ltd
Public Summary: We hypothesise that increased bacterial exo-enzyme activity in a high CO2 ocean will decrease the carbon pump to the deep ocean. Acidification of the ocean is a result of increasing CO2 in the atmosphere that is now affecting the ecology of the ocean. However, the impact of high CO2 in the oceans on marine bacteria is not well understood or researched, even though marine bacteria play a fundamental role in cycling nutrients and transferring carbon to the deep ocean. We propose to investigate this using both CO2 manipulation incubations, and also natural areas of high CO2 in the Bay of Plenty. This combined approach is a powerful one that overcomes the limitation of short-term laboratory CO2 incubations, and will enable prediction and modelling of changes in bacterial activity in the future ocean.
Total Awarded: $756,522
Duration: 3
Host: National Institute of Water and Atmospheric Research
Contact Person: Dr CS Law
Panel: EEB
Project ID: 10-NIW-009
Fund Type: Marsden Fund
Category: Standard
Year Awarded: 2010
Title: Bacterial geo-thermometer: A new, precise indicator of climate change
Recipient(s): Dr MJ Vandergoes | PI | GNS Science
Dr AC Dieffenbacher-Krall | AI | University of Maine
Prof RM Newnham | AI | Victoria University of Wellington
Prof L Schwark | AI | University of Kiel
Dr K Zink | AI | GNS Science
Public Summary: Accurate reconstructions of past temperature change from Southern Hemisphere locations are scarce and are needed urgently to provide a long term context to current climate change, test leading hypotheses about possible causal mechanisms, and to constrain models used to predict future climate. Recent progress has been made with new models for quantitative climate reconstruction using biological indicators including pollen and insect remains. A key question is whether these reconstructions can be validated and their precision improved by applying additional quantitative techniques. A recently developed method using microbial membrane lipids in lake sediments as a molecular paleothermometer has the potential to enable more precise paleotemperature reconstructions than have been previously achieved in New Zealand. The proposed research will establish and apply this technique to New Zealand paleoclimate sites that span the last glaciation. This innovative approach will increase the precision of New Zealand’s paleoclimate estimates and address issues relating to the timing, magnitude and variability of last glaciation climate change. The results will allow for better integration of terrestrial and marine reconstructions and more robust climate correlation between the northern and the southern hemispheres.
Total Awarded: $686,957
Duration: 3
Host: GNS Science
Contact Person: Dr MJ Vandergoes
Panel: ESA
Project ID: 10-GNS-001
Fund Type: Marsden Fund
Category: Standard
Year Awarded: 2009
Title: Bacterial protection against phage infections
Recipient(s): Dr PC Fineran | PI | University of Otago
Professor GPC Salmond | AI | University of Cambridge
Public Summary: Phage are viruses that infect bacteria. They are the most abundant biological entities on the planet and are involved in the lysis of bacteria at enormous rates, driving bacterial evolution and biogeochemical cycles. How bacteria evade phage is a key, under-examined field. We will test if there is a relationship between the abundant bacterial toxin-antitoxin (TA) and abortive infection (Abi) loci as 'cell suicide' modules adapted for viral evasion. Their mechanisms used to sense and abort phage replication and protect bacterial populations will be investigated, leading to a greater understanding of the most frequent predator-prey interaction.
Total Awarded: $669,333
Duration: 3
Host: University of Otago
Contact Person: Dr PC Fineran
Panel: CMP
Project ID: 09-UOO-177
Fund Type: Marsden Fund
Category: Standard
Year Awarded: 2013
Title: BAFfled: how does orf virus defeat the BAF cellular defence mechanism?
Recipient(s): Professor AA Mercer | PI | University of Otago
Professor P Traktman | AI | Medical College of Wisconsin
Public Summary: It was shown recently that a eukaryotic protein, BAF, is part of the cell’s innate defences against viral infection. Most poxviruses defeat this defence by using a viral kinase to phosphorylate BAF. However, we have identified some poxviruses that lack this kinase. How then do these viruses counteract BAF? We will identify the factor these viruses use to defeat BAF and characterise its mechanism of action. This project will define a new mechanism by which viruses evade cellular defences and shine new light on the activities of BAF, an essential cellular factor.
Total Awarded: $800,000
Duration: 3
Host: University of Otago
Contact Person: Professor AA Mercer
Panel: CMP
Project ID: 13-UOO-233