Breakthrough(20)

Lokke’s expression became concerned. “Go on.”

“We’ve got a couple researchers there in the Halley camp; got thrown out of bed in the middle of the night. The next morning they went out to take a look. They found a crack, a big crack.”

Lokke held her breath. “How big?”

“Miles. “

Her next question was very deliberate. “Where?”

Haines inhaled slowly. “Beyond the shelf, on land. About ten miles in.”

She stared at him for a long time then finally leaned back in her chair. “Jesus.” Haines sat silently while it sank in. Lokke looked at a large map on the wall of the Earth’s southern hemisphere. “How bad is the slide?” she asked.

“Fourteen feet.”

“Oh my god!” She shook her head in disbelief. “Has this been verified?”

“Yes. We had a plane fly down from McMurdo. They ran the length and measured 43 miles, most of it on land. The ground team verified the slide depth nearest the camp. It looks pretty consistent from the air. It was strong enough to set off all the seismic instruments on the continent.”

Lokke knew what that meant. Even though the Antarctic was a rather small land mass, all of the earthquake sensors being triggered meant the shift was powerful, very powerful. “Let’s get a large team and some equipment out there right away. We need to get all the measurements and verify with a ground penetrating scan before we communicate this out.”

Haines nodded.

“In the meantime, I’ll start getting people together.”





13





Chris replaced the camcorder’s tape with a new one and handed the old tape to Alison who applied a small label on its back displaying a tiny barcode. She then passed it under an optical scanner which recorded the number and brought up a small field on the computer’s monitor to add notes. She typed a short summary of the exercise they had just completed and hit enter, adding the entry to the database.

“Ready for the next one?” Chris asked.

“Let’s give them a break,” she said looking at Dirk and Sally. She glanced down at her watch. “It’s almost feeding time and I’m sure we’ve got a-”

Food now. Sally’s words came over the speakers.

Alison smiled. She leaned over and typed Yes, food now and clicked the translate button. She turned around and picked up the phone to call the aquarium’s feeding team.

Thank Alison, came the mechanical reply.

Alison suddenly froze. She dropped the phone and turned around to look at the tank. Both Chris and Lee had done the same. After a long silence, they all turned toward each other.

“What did she say?” asked Alison.

Chris leaned past Lee and looked at the monitor screen. “Did Sally just say ALISON?”

Lee scooted forward and studied the screen. “Uh…she sure did.” He raised one of his eyebrows. “How the hell did she do that?” Lee typed several words on his computer and hit the enter key. Another window opened, displaying the transaction log files and hundreds of lines of cryptic text. He identified the last line translated and looked at the attribute signature of the word Alison. He then scrolled down examining the log details for all translations over the last two days. “Aha!” he said triumphantly, but suddenly leaned back in his seat with another thought. “Oh wow!”

“What is it?” Alison pressed.

He did not answer immediately. “Wow that is really something else.”

“What already?!”

“Context. IMIS translated for context.” He almost seemed to be talking to himself. He quickly snapped out of it and looked at Alison and Chris. “You see, the design of IMIS’ artificial intelligence, part of what allows it to learn, is trying to identify context. This is basically the relationship between multiple variables to achieve what it believes to be the best accuracy.”

Chris frowned. “This is starting to sound Greek.”

“Look,” Lee said, starting to get excited. “Let’s say IMIS is trying to figure out the word hello. If it thought it identified the right dolphin sound for hello but that sound was emitted when the dolphins were moving away, the context of that interaction would show that translation to be wrong.”

Alison nodded. “Of course.”

“Obviously that’s an over simplified example. What IMIS just did was apply that context on BOTH sides of the translation. Look at this!” He pointed to the translation logs on the screen. “What Sally said was not Thank Alison, she actually said Thank girl. But IMIS not only identified and translated girl successfully but it also used its context algorithm to recognize that Alison is the only female among us; probably based on her computer account since she has been typing some of the words to Dirk and Sally. Therefore IMIS recognized that the message was meant for the only girl and substituted Alison’s name. Hell it even capitalized her name.”