Fourth Annual Three Minute Thesis Competition

[Speaker: Graham Hammill] Our next presenter is Joëlle Carota from the Department of Romance Languages and Literatures, and the title of her 3MT presentation is "The Venezuelan Community of Pescara (Italy): A Sociolinguistic Analysis of Spanish in Contact With Italian." And one thing I'll just share about you is that, not only is Joëlle an expert in romance languages, but she's also learning American Sign Language. So, ready, set, pitch.

[Speaker: Joëlle Carota, speaking Spanish] As you might have noticed, yes, that was Spanish and it's actually very common for bi- and multilinguals like myself to use all languages available in their linguistic repertoire and to switch from one language to another in their daily interactions like I just did.

In my dissertation project, I seek to shed some light on how the bilingual members of the Italo-Venezuelan community of Pescara—my hometown in Italy—use linguistic variation to create, shape and portray their social persona.

This is actually the first quantitative study that wants to understand how linguistic variation and manipulation of Spanish and Italian affects processes such as language maintenance and or language loss in a foreign context. Let us take, as an example, these set of two identical words, example number two, that translate as Easter in English. They would be pronounced in two different ways. In Italian: Pasqua, and in Venezuelan Spanish: Paqua. Thereby the speaker is left with two different options. The speaker will choose one of the two options depending on the message they want to convey to their audience. And we actually really want to know what that message is, as well as what is the answer that the speaker has for questions such as, are Italo-Venezuelans integrated into the Italian community? What do they think of the local school system? How are they approaching local lifestyle in culture and language?

You see, knowing the answer to these questions is very important for us, so that we can first validate the very existence of this community. And also, we can help them in integrating successfully in the new community. By observing and studying natural language production, we want to create comprehensive linguistic and educational policies that could be implemented both at a national- and local-level, and that take into account the linguistic diversity that has existed in the Italian school system for quite some time now.

Also, I would like to create Spanish as a heritage language programs and Italian as a second language classes that targets specifically these communities, so that the younger generation do not lose their heritage, language and culture.

Think how much better things could be if this immigrant communities felt compelled to enhance the environment in which they live. But we cannot help them doing so without first getting to know them. And this is what motivates my research. Thank you very much.

[Audience applauding.]

[Speaker: Graham Hammill] Our sixth presenter is Poonam Choudhary from the Department of Medicine. The title of her presentation is "Advanced Imaging for Pre-Symptomatic Diagnosis of Krabbe Disease." Not only is she an expert in medical sciences, but she also has an interest in fashion design and can speak, read and write three languages fluently, which is quite impressive. So, you ready, set, pitch.  

[Speaker: Poonam Choudhary] Good afternoon, everybody. Let me start with an assumption that if you are from Buffalo, and an avid football follower, you might have heard of Krabbe disease. Have you? Well, if you haven't, let me inform you about it.

Krabbe is a rare neurological disorder, which becomes fatal within two to five years of life. And being genetic in nature, it is currently impossible to predict and prevent it from happening. Now, if you ask me about the two most important factors in treating a disease, my answer would be the accuracy of the diagnosis and its timely treatment.

Shown on the left hand side is a newborn screening, a process of testing blood samples for newborns for Krabbe disease. But there are some issues with this process. For one, there's 25 percent chance of detecting false positives or false negatives. Meaning Krabbe disease can either go undetected, or it can be falsely diagnosed. Second, not being federally-mandated procedure, it is only applicable in just few states. And New York is actually one of them. So when a child gets diagnosed, or misdiagnosed with Krabbe, that child has to go through an excruciating stem cell therapy, which comes with heavy side effects. Nobody, nobody likes to be treated for what they don't have, especially infants. The physical and emotional toll on the child and the parents is just unimaginable.

My research shows that by using advanced imagine technology of PET and MRI, it is indeed possible to detect pre-symptomatic biomarker of Krabbe disease in 15 days old mice. And by the word “pre-symptomatic,” I mean the time before the physical and clinical symptoms appear, which is very crucial period to nip the disease in the bud.

In future, we will translate this animal study in humans. And if we are successful, then we will be in a better position to recommend advanced imaging, along with newborn screening across all medical centers in the country.

To conclude, for those of you who are still wondering about the connection between football and Krabbe I mentioned in the beginning, for them, my suggestion is to go home and do some research about it. And here's a hint. The Buffalo Bills' quarterback were to Buffalo to Super Bowl four years in a row. Thank you.

[Audience applauding.]

[Speaker: Graham Hammill] The tenth presenter for today is Zeinab Farhat, who's a PhD student in epidemiology and environmental health in the School of Public Health and Health Professions. The title of her 3MT presentation is "Garlic's Role as an Antioxidant in Cancer." She enjoys going to concerts, cooking, baking and doing puzzles, and she's both an avid fan of the New York Giants, which is okay, and iced coffee. So, ready, set, pitch.

[Speaker: Zeinab Farhat] Has anyone ever told you to eat garlic when you're sick or to use garlic to treat infection? That's because garlic is a superfood that has been used since ancient times. But only recently, we'd begun to understand its wide range of health benefits.

Garlic contains active compounds such as amino acids, vitamins, minerals, and over 33 sulfur compounds, which we can attribute those health benefits to. One such health benefit is its anti-cancer activity, which is the focus of my dissertation project.

About one-third of all cancer deaths in the United States could be avoided to appropriate dietary changes. And garlic is a cancer preventing food that has the ability to trap, block and suppress cancer causing agents from leading to cancer. Garlic can act by capturing molecules known as free radicals that can roam around in the cell, causing damage to our DNA, and ultimately leading to cancer. Or it can act by directly killing cancer cells. These free radicals are angry molecules that can arise from sources, such as cigarette use, alcohol use, air pollution exposure, stress and drug use. If we have too many of these free radicals roaming around in the cell, this can lead to a dangerous phenomenon known as oxidative stress. This phenomenon can be counteracted by antioxidant like garlic, because they can capture these free radicals and prevent them from causing DNA damage.

The first part of my dissertation focuses on utilizing data from a large, nationwide human study that collected information on people's garlic intake and supplement use and follow them for a period of time to see whether or not they got cancer or they died from cancer. What we found was that garlic supplement use for three or more years could prevent against any cancer development.

However, the challenge with epidemiological or human studies is that the biologically active compounds in garlic could differ depending on the type of garlic form that's eaten, or how its prepared. So I set out to test the antioxidant and anti-cancer activity of several garlic forms, including fresh garlic cooked garlic, garlic powder, and two commercial available garlic supplements. What we found was that Garlicin, which is a garlic supplement and contains predominantly garlic powder, had the ability to have high antioxidant activity and had the ability to kill cancer cells in lung cancer cell lines.

The next part of my project was to administer this Garlicin supplement to a group of healthy volunteers here at UB in a clinical trial and measure their antioxidant activity in the blood over a period of six weeks. Ultimately, we hope to translate these findings to cancer patients and promote garlic as a part of the regular healthy diet. Thank you.

[Audience applauding.]

[Speaker: Graham Hammill] I have to admit, that presentation made me a little hungry.

[Speaker: Graham Hammill] The next presenter is Farshad Ghanei, who's a PhD student in the Department of Computer Science and Engineering, and the title of his presentation is "Energy Awareness." Not only does he have a third degree black belt, which is very impressive, but there's this sort of coincidence that he's been teaching at UB for six years, and also has had a small pet snail for six years. I don't know if there's a connection there, but at least it's very interesting. So if you're ready, set, pitch.

[Speaker: Farshad Ghanei] How often does your phone run out of battery? Have you thought whether your devices run as efficiently as they can? And finally, are you interested in extending the runtime of your phones or drones? I have good news for you. My research can help your mobile devices run more efficiently.

We live in a world surrounded by electronic devices, and we use them every single day. Many of them rely on a battery and they have one thing in common, their energy is limited. Some of them may run for a few hours before they run out, such as your phones or laptops. And some of them may only work for less than an hour. And for the most part, it's not even easy, or in cases possible, to charge them quickly and conveniently. So we, the developers and programmers, should be very, very careful how to design and program these devices, so they use their limited energy efficiently.

My dissertation introduces energy awareness in battery-powered mobile devices. I study mobile systems, specifically phones and drones, and introduce ways so that we can make them more energy efficient. I achieve this in three steps.

Step one, I introduce software changes that allow the system to track its energy use. Think of your phones and applications. Your system needs to know how much energy is used by each application, so that it can report it correctly, and close them if needed. And it should know this accurately. My research helped achieve this by modifications in the underlying operating system.

Step two, I use a specialized hardware to measure the actual energy consumption of a device so that we can use it for better planning. Think of a drone covering an area looking for an object or a person. The trajectory that you plan for this flight can hugely impact its energy and runtime. And we have shown that we can improve this by 15 to 20 percent.

Step three, I introduce energy awareness to the next level. I let the device decide by itself and then show how task specific energy awareness can lead to more utilization and efficiency. So that the device, at runtime, can make the most energy efficient decision. This is what we call energy awareness, where we study what impacts the energy of battery powered systems and use this information for better modeling, planning and decision making.

My dissertation can help the designers, developers and programmers program our devices, such that they run more efficiently and for longer periods of time. Thank you.

[Audience applauding.]

[Speaker: Graham Hammill] Our final presenter for today is Nagashri Lakshminara... I'm sorry, Nagashri Lakshminarayana, I apologize for that. She's a student in the Department of Computer Science and Engineering in the School of Engineering and Applied Sciences. The title of her presentation is "Facial Expression Recognition: A Step Towards Emotionally Intelligent Technology.” Besides being an avid hiker, she also is an avid game player when she's with friends. So, ready, set, pitch.

[Speaker: Nagashri Lakshminarayana] Do you know what your face can say about you? Depending on the situation, you're either smiling, frowning or showing some other form of expression. Your face has a way of letting people know how you feel without using words.

Here's an interesting story. Dr. Paul Ekman, a renowned social psychologist found something unusual in a specific video of a patient. The patient was trying to convince her counselors that she no longer had suicidal instinct. Everything seemed believable until he went through the video one frame at a time. Through a series of hidden emotions, he found a small, but very strong glimpse of distress that the patient was trying to hide. This wouldn't have been detected by a naked eye. This was the beginning of facial expression recognition.

For humans to annotate this expression is an extremely time consuming and a cumbersome task. Instead, we can use computers to augment the effort and make the processing faster.

My PhD thesis uses artificial intelligence to help computers perceive emotions the way that humans naturally do. With the image as an input, my algorithm learns to predict minuscule muscle movements on face, like an eyebrow rise, or a lip corner puller, to name a few. Between the input and the output are a series of complex functional blocks that are inspired by human intelligence, paying more attention to the salient regions on face, like the region around eyes and mouth that are defined on anatomical basis. Each of these functions model increasingly complex features starting from edges and blobs in images, all the way to human perceivable forms like eyes and nose.

With no explicit programming, my algorithms can automatically leverage human expertise and match their predictions. Further, they can be used in scenarios where such expertise is unavailable or expensive.

In the past few decades, there's been a rapid shift in how we interact with technology from using computers merely for calculations to social computing. Therefore, now more than ever, there's a greater need for technology to be able to support human cognitive capabilities.

Looking ahead, this could benefit many applications, like automatically detecting symptoms of pain and depression in geriatric patients who are unable to verbalize their feelings or to detect and alert the automobile users to their emotional states.

My PhD thesis is one step towards creating technology that can recognize and respond to human emotions effectively, thereby creating more meaningful human computer interactions. Thank you.

[Audience applauding.]

[Speaker: Graham Hammill] I hope everybody here enjoyed the presentations as much as I did. It's always so impressive to hear the amazing research that is happening here at the university. I'd like to give a round, I'd like to ask everybody to give a round of applause to everyone who presented here today for all of the excellent performances that we saw. And this brings an end to this year's Three Minute Thesis Competition. I can't wait to see who the winners are and I'm glad I don't have to choose.

[Speaker: Graham Hammill] The first presenter is Seyed Omid Sajedi­—please come up—who is in the Department of Civil Structural and Environmental Engineering, and the 3MT presentation title is "Artificially Intelligent Systems for Rapid Post-Earthquake Inspections." Ready, set, pitch.

[Speaker: Seyed Omid Sajedi] Earthquakes are terrible things. In my home country, Iran, in a city with 90,000 residents, more than two-thirds of the population lost their lives after a massive earthquake. Unfortunately, human casualties are just the tip of the iceberg. After earthquakes, most buildings and bridges cannot be used before they are properly inspected. Therefore, the economic and social consequences could be very severe and potentially escalate beyond the seismic zone.

A recent report showed that an earthquake in California, which supplies 30 percent of the food in the whole United States, could cost over $200 billion. Just imagine the consequences of shutting down this interchange for a few days, it's going to be bad. We need to quickly know if our civil infrastructures are safe, so people can get back to their normal lives as soon as possible.

Having said that, information on a structural safety is crucial, not to just rescue people, but also to minimize these economic losses. Unfortunately, human inspection are not a good option. They require a lot of time and real money. And if any of these resources are available, are you willing to go inside the building or under a bridge like this that may collapse any minute? I know I'm not. And that's the reason my research at UB is focused on developing systems that can predict damage in real time.

Like a human heart monitor, structural vibrations tell us a lot about the condition of our buildings. With the help of structural dynamics and artificial intelligence, we are trying to translate these vibrations into damage in just a few seconds. My research is dedicated to making damage detectors that can work reliably even in very noisy environments.

Therefore, our artificial neural networks are trained and tested for thousands of earthquake simulations, so they can predict various types of damage. As a result, by placing a few accelerometers here and there on the bridge or a building, you can quickly predict, not only the locations of damage, but also their severities.

Sadly, history and science have taught us that earthquakes are inevitable. But disasters don't have to be. With the help of my research, we are trying to train robot doctors for our cities. Then they are giving us valuable information so we can take proper actions immediately after the earthquakes. The efforts of me and my colleagues will take us several steps closer to faster recovery and building more resilient communities against these natural disasters. Thank you so much everyone and stay safe.

[Audience applauding.]

[Speaker: Graham Hammill] Thank you, Seyed.

[Speaker: Graham Hammill] The second presenter is Emily Sekera, who's a student in the Department of Chemistry. Emily's 3MT presentation title is "More Than a Gut Feeling: Finding a Biomarker for Autism and Spectrum Disorder". One of the things about Emily that I'll just share with you is that, not only is she an expert in chemistry, but she's also a lover of wine and goes to wine tastings quite often and apparently has also taken wine tasting classes. So, that's really terrific. So if you're ready, ready, set, pitch.

[Speaker: Emily Sekera] I'd like you to imagine for a moment that you are a new parent. You're excited, you're nervous and a little bit sleep deprived. As you watch your child grow, you notice that they're not really looking you in the eye, they're not babbling, but maybe you're just overreacting. As time goes on, though, they're not playing with the other children and they're not talking. All right, it's time to see a doctor. And after what feels like countless doctors' appointments, they finally come back and tell you, "All right, your child probably has autism… Probably? What do you mean probably? Shouldn't you be giving me a definitive answer?"

Unfortunately, that's the reality that those with autism spectrum disorders currently have to face: a probably. We're just missing one more puzzle piece.

Everyone deserves a definitive diagnosis and not just a probably, yet despite its increasing incidence from one in 150 to one in 59 in just a decade, we still have no answer. But that's where I come in.

I'm looking for a way to diagnose autism in non-invasive samples, like urine and feces. And I found something: It's what causes your feces to be brown and your urine to be yellow. It's name? Stercobilin, and you can see it right in our puzzle pieces to the left. And it makes some sense too. Stercobilin is made in our guts. And the guts and the bacteria in them are different in those with autism than they are in healthy controls.

So, where are we at now? I've been doing some testing in animal models. In the first, we use an animal model using a mouse with Timothy Syndrome, which, when in humans, over 80% of those who live long enough tend to develop ASD. And what we found was astonishing. In just 14 pairs of mice, we saw a depletion of 48 percent in the autistic mice. And this is with a significance to us that could be used in the clinic.

But it wasn't enough for me just yet. So we've moved into a larger model. This time, rats. But instead of Timothy Syndrome, they had Fragile X. So, Fragile X is the one and only genetic cause we really know for autism at the moment. And currently with 13 pairs of mice, we are seeing a depletion of Stercobilin in of 38 percent, but we're not quite ready for the clinic yet. For now, that is.

So, in the last few months of my PhD, that's what I'm planning to do. I'm planning to bring diagnosis to clinical trial. As I said, everyone deserves a definitive diagnosis. And it's time we diagnose autism. Thank you.

[Audience applauding.]

[Speaker: Graham Hammill] Our eighth presenter today is Xiaoxiao Zhang, who's a PhD student in the Department of Oral Biology in the School of Dental Medicine. The title of her presentation is "Preventing Bone Loss with Fewer Side Effects." One of the things that Xiaoxiao says about herself is that for fun, she likes hiking, salsa dancing, and weightlifting. That's very impressive. So, if you're ready, set, pitch.

[Speaker: Xiaoxiao Zhang] It's hard to notice, but our bones are changing all the time. While I'm talking, my body is breaking down old bone and building up new bone continuously. These two processes are conducted by bone destroying cells and bone building cells. Like our stomachs produce stomach acid to digest food, bone destroyers produce Enzyme K to digest the bone. In healthy adults, bone mass is stable because bone destroyers and bone builders are constantly talking to each other by sending and receiving signals. However, when people age, bone destroyers may go out of control, as you can see on the screen. They produce too much Enzyme K, which are like bombs blowing up healthy bone. This is how osteoporosis occurs.

Osteoporosis puts people at great risk for fractures. In the U.S., half of women and a quarter of men over 50 will experience an osteoporosis related fracture. However, commonly used osteoporosis drugs are low efficiency, because they aim at killing bone destroyers. Indeed, killing them can prevent bone loss, but it also disrupts their communication with bone builders. You can imagine when bone destroyers keep dying, bone builders cannot get enough signals from them. As a result, less new bone is built. So how do we get without this terrible side effect?

Instead of killing bone destroyers, what if we just disarm their bombs, Enzyme K?

Fortunately, a substance called heparin sulfate has been shown to bind to Enzyme K and inhibit its activity. However, there are so many different kinds of heparin sulfate. My research is to identify which heparin sulfate interacts with Enzyme K the best.

To achieve this, we use a very cool technique called microarray. It allowed us to compare the binding affinity of Enzyme K to 52 heparin sulfate in a single experiment. We have already identified a few candidates. Next, we will test them on mice.

Our ultimate goal is to invent a heparin sulfate based drug that inhibits Enzyme K, and at the same time, brings fewer side effects than traditional osteoporosis drugs. This research is novel, because there's no drug targeting Enzyme K so far. Hopefully, in the near future, there will be no osteoporosis anymore. And people can all have strong bones at old age. Thank you.

[Audience applauding]